A machine learning method for acute hypotensive episodes prediction using only non-invasive parameters.

BACKGROUND AND OBJECTIVES: Accurate prediction of acute hypotensive episodes (AHE) is fundamentally important for timely and appropriate clinical decision-making, as it can provide medical professionals with sufficient time to accurately select more efficient therapeutic interventions for each specific condition. However, existing methods are invasive, easily affected by artifacts and can be difficult to acquire in a pre-hospital setting. METHODS: In this study, 1055 patients' records were extracted from the Multiparameter Intelligent Monitoring in Intensive Care II database (MIMIC II), comprising of 388 AHE records and 667 non-AHE records. Six commonly used machine learning algorithms were selected and used to develop an AHE prediction model based on features extracted from seven types of non-invasive physiological parameters. RESULTS: The optimal observation window and prediction gap were selected as 300 minutes and 60 minutes, respectively. For GBDT, XGB and AdaBoost, the optimal feature subsets contained only 39% of the overall features. An ensemble prediction model was developed using the voting method to achieve a more robust performance with an accuracy (ACC) of 0.822 and area under the receiver operating characteristic curve (AUC) of 0.878. CONCLUSION: A novel machine learning method that uses only noninvasive physiological parameters offers a promising solution for easy and prompt AHE prediction in widespread scenario applications, including pre-hospital and in-hospital care.

A new method for identifying the acute respiratory distress syndrome disease based on noninvasive physiological parameters.

Early diagnosis and prevention play a crucial role in the treatment of patients with ARDS. The definition of ARDS requires an arterial blood gas to define the ratio of partial pressure of arterial oxygen to fraction of inspired oxygen (PaO2/FiO2 ratio). However, many patients with ARDS do not have a blood gas measured, which may result in under-diagnosis of the condition. Using data from MIMIC-III Database, we propose an algorithm based on patient non-invasive physiological parameters to estimate P/F levels to aid in the diagnosis of ARDS disease. The machine learning algorithm was combined with the filter feature selection method to study the correlation of various noninvasive parameters from patients to identify the ARDS disease. Cross-validation techniques are used to verify the performance of algorithms for different feature subsets. XGBoost using the optimal feature subset had the best performance of ARDS identification with the sensitivity of 84.03%, the specificity of 87.75% and the AUC of 0.9128. For the four machine learning algorithms, reducing a certain number of features, AUC can still above 0.8. Compared to Rice Linear Model, this method has the advantages of high reliability and continually monitoring the development of patients with ARDS.

[Research on algorithms for identifying the severity of acute respiratory distress syndrome patients based on noninvasive parameters].

Acute respiratory distress syndrome (ARDS) is a serious threat to human life and health disease, with acute onset and high mortality. The current diagnosis of the disease depends on blood gas analysis results, while calculating the oxygenation index. However, blood gas analysis is an invasive operation, and can't continuously monitor the development of the disease. In response to the above problems, in this study, we proposed a new algorithm for identifying the severity of ARDS disease. Based on a variety of non-invasive physiological parameters of patients, combined with feature selection techniques, this paper sorts the importance of various physiological parameters. The cross-validation technique was used to evaluate the identification performance. The classification results of four supervised learning algorithms using neural network, logistic regression, AdaBoost and Bagging were compared under different feature subsets. The optimal feature subset and classification algorithm are comprehensively selected by the sensitivity, specificity, accuracy and area under curve (AUC) of different algorithms under different feature subsets. We use four supervised learning algorithms to distinguish the severity of ARDS (P/F ≤ 300). The performance of the algorithm is evaluated according to AUC. When AdaBoost uses 20 features, AUC = 0.832 1, the accuracy is 74.82%, and the optimal AUC is obtained. The performance of the algorithm is evaluated according to the number of features. When using 2 features, Bagging has AUC = 0.819 4 and the accuracy is 73.01%. Compared with traditional methods, this method has the advantage of continuously monitoring the development of patients with ARDS and providing medical staff with auxiliary diagnosis suggestions.

A method to differentiate between ventricular fibrillation and asystole during chest compressions using artifact-corrupted ECG alone.

In recent years, numerous adaptive filtering techniques have been developed to suppress the chest compression (CC) artifact for reliable analysis of the electrocardiogram (ECG) rhythm without CC interruption. Unfortunately, the result of rhythm diagnosis during CCs is still unsatisfactory in many studies. The misclassification between corrupted asystole (ASY) and corrupted ventricular fibrillation (VF) is generally regarded as one of the major reasons for the poor performance of reported methods. In order to improve the diagnosis of VF/ASY corrupted by CCs, a novel method combining a least mean-square (LMS) filter and an amplitude spectrum area (AMSA) analysis was developed based only on the analysis of the surface of the corrupted ECG episode. This method was tested on 253 VF and 160 ASY ECG samples from subjects who experienced cardiac arrest using a porcine model and was compared with six other algorithms. The validation results indicated that this method, which yielded a satisfactory result with a sensitivity of 93.3%, a specificity of 96.3% and an accuracy of 94.8%, is superior to the other reported techniques. After improvement using the human ECG records in real cardiopulmonary resuscitation (CPR) scenarios, the algorithm is promising for corrupted VF/ASY detection with no hardware alterations in clinical practice.

[Research on malignant arrhythmia detection algorithm using neural network optimized by genetic algorithm].

Detection and classification of malignant arrhythmia are key tasks of automated external defibrillators. In this paper, 21 metrics extracted from existing algorithms were studied by retrospective analysis. Based on these metrics, a back propagation neural network optimized by genetic algorithm was constructed. A total of 1,343 electrocardiogram samples were included in the analysis. The results of the experiments indicated that this network had a good performance in classification of sinus rhythm, ventricular fibrillation, ventricular tachycardia and asystole. The balanced accuracy on test dataset reached up to 99.06%. It illustrates that our proposed detection algorithm is obviously superior to existing algorithms. The application of the algorithm in the automated external defibrillators will further improve the reliability of rhythm analysis before defibrillation and ultimately improve the survival rate of cardiac arrest.

[Development of a microenvironment test chamber for airborne microbe research].

One of the most important environmental cleanliness indicators is airborne microbe. However, the particularity of clean operating environment and controlled experimental environment often leads to the limitation of the airborne microbe research. This paper designed and implemented a microenvironment test chamber for airborne microbe research in normal test conditions. Numerical simulation by Fluent showed that airborne microbes were evenly dispersed in the upper part of test chamber, and had a bottom-up concentration growth distribution. According to the simulation results, the verification experiment was carried out by selecting 5 sampling points in different space positions in the test chamber. Experimental results showed that average particle concentrations of all sampling points reached 10 7 counts/m 3 after 5 minutes' distributing of Staphylococcus aureus, and all sampling points showed the accordant mapping of concentration distribution. The concentration of airborne microbe in the upper chamber was slightly higher than that in the middle chamber, and that was also slightly higher than that in the bottom chamber. It is consistent with the results of numerical simulation, and it proves that the system can be well used for airborne microbe research.

A new method without reference channels used for ventricular fibrillation detection during cardiopulmonary resuscitation.

Ventricular fibrillation (VF) is observed as the initial rhythm in the majority of patients suffering from sudden cardiac arrest. It is vitally important to accurately recognize the initial VF rhythm and then implement electrical defibrillation. However, artifacts produced by chest compression during cardiopulmonary resuscitation (CPR) make the VF detection algorithms utilized by current automated external defibrillators (AEDs) unreliable. CPR must be traditionally interrupted for a reliable diagnosis. However, interruptions in chest compression have a deleterious effect on the success of defibrillation. The elimination of the CPR artifacts would enable compressions to continue during AED VF detection and thereby increase the likelihood of resuscitation success. We have estimated a model of this artifact by adaptively incorporating noise-assisted multivariate empirical mode decomposition (NA-MEMD) and least mean squares (LMS) and then removing the artifact from the corrupted ECGs. The simulation experiment indicated that the CPR artifact could be accurately modeled without any reference channels. We constructed a BP neural network to evaluate the results. A total of 372 VF and 645 normal sinus rhythm (SR) ECG samples were included in the analysis, and 24 CPR artifact signals were used to construct corrupted ECGs. The results indicated that at different SNR levels ranging from 0 to -12 dB, the sensitivity and specificity were always above 95 and 80 %, respectively.

[A New Algorithmic Method to Detect Ventricular Fibrillation Using Electrocardiogram Signals During Cardiopulmonary Resuscitation by Artificial Pressing].

On account of the mechanical disturbance of external chest pressing to electrocardiogram(ECG)signal,the ECG rhythm cannot be identified reliably during the cardio-pulmonary resuscitation period.Whereas the possibility of successful resuscitation will be lowered due to interrupted external chest pressing,a new filtering algorithm,enhanced leastmean-square(eLMS)algorithm,was proposed and developed in our laboratory.The algorithm can filter the disturbance of external chest pressing without the support of hardware reference signal and correctly identify ventricular fibrillation(VF)rhythm and normal sinus rhythm in case of uninterrupted external chest pressing.Without other reference signals,this algorithm realizes filtering only through the interrupted electrocardiograma(cECG)signal.It was verified with ECG signal and disturbance signal under different signal to noise ratios and contrasted with other mature algorithms.The verification results showed that the identification effect of eLMS was superior to those of others under different signal to noise ratios.Furthermore,ECG rhythm can be correctly identified only through cECG signal.This algorithm not only reduces the research and development(R&D)costs of automated external defibrillator but also raises the identification accuracy of ECG rhythm and the possibility of successful resuscitation.

[Adaptive Cardiopulmonary Resuscitation Artifacts Elimination Algorithm Based on Empirical Mode Decomposition and Independent Component Analysis].

Artifacts produced by chest compression during cardiopulmonary resuscitation(CPR)seriously affect the reliability of shockable rhythm detection algorithms.In this paper,we proposed an adaptive CPR artifacts elimination algorithm without needing any reference channels.The clean electrocardiogram(ECG)signals can be extracted from the corrupted ECG signals by incorporating empirical mode decomposition(EMD)and independent component analysis(ICA).For evaluating the performance of the proposed algorithm,a back propagation neural network was constructed to implement the shockable rhythm detection.A total of 1 484 corrupted ECG samples collected from pigs were included in the analysis.The results of the experiments indicated that this method would greatly reduce the effects of the CPR artifacts and thereby increase the accuracy of the shockable rhythm detection algorithm.

[Analysis and Optimization of the Temperature Retard between Sample and Air Based on Nucleic Acid Amplification System Heated by Air].

It is the main method for amplifying the specific gene to use the nucleic acid amplification system to accomplish polymerase chain reaction(PCR).The temperature retard between heat source and sample exists in the heating and cooling progresses of most nucleic acid amplification system.The retard would result in the problem that the sample would take a long time to reach the set temperature and the problem would reduce the speed of integrate reaction.Non-specific products would be created in the process of amplification when the sample cannot reach the set temperature within a certainly time and the amplified efficiency would be reduced.A miniaturization nucleic acid amplification system heated by air was designed in this study according to the principle of air-heated nucleic acid amplification system and the characteristics of the PCR instrument Smart-cycler.The heat transfer process was analyzed and the heat transfer time was calculated.The actual temperature was measured in real time,and the temperature curves were fitted.The heating time was chosen by analysis results and data fitting and the air temperature was changed,while the sample temperature was recorded.The retard between sample and air was optimized by choosing the best curve of sample temperature.The temperature retard between sample and air was reduced sharply and the required time of integrate progress is shortened to 50%.We confirmed from the amplification experiment of Listeria monocytogenes that the improved system could complete 3cycles within 4minutes,and the amplification effect was good.The amplification speed and effect could be improved effectively by optimizing the delay between sample and air.

Closed-loop controller for chest compressions based on coronary perfusion pressure: a computer simulation study.

High-quality cardiopulmonary resuscitation contributes to cardiac arrest survival. The traditional chest compression (CC) standard, which neglects individual differences, uses unified standards for compression depth and compression rate in practice. In this study, an effective and personalized CC method for automatic mechanical compression devices is provided. We rebuild Charles F. Babbs' human circulation model with a coronary perfusion pressure (CPP) simulation module and propose a closed-loop controller based on a fuzzy control algorithm for CCs, which adjusts the CC depth according to the CPP. Compared with a traditional proportion-integration-differentiation (PID) controller, the performance of the fuzzy controller is evaluated in computer simulation studies. The simulation results demonstrate that the fuzzy closed-loop controller results in shorter regulation time, fewer oscillations and smaller overshoot than traditional PID controllers and outperforms the traditional PID controller for CPP regulation and maintenance.

A mechanical chest compressor closed-loop controller with an effective trade-off between blood flow improvement and ribs fracture reduction.

Chest compression (CC) is a significant emergency medical procedure for maintaining circulation during cardiac arrest. Although CC produces the necessary blood flow for patients with heart arrest, improperly deep CC will contribute significantly to the risk of chest injury. In this paper, an optimal CC closed-loop controller for a mechanical chest compressor (OCC-MCC) was developed to provide an effective trade-off between the benefit of improved blood perfusion and the risk of ribs fracture. The trade-off performance of the OCC-MCC during real automatic mechanical CCs was evaluated by comparing the OCC-MCC and the traditional mechanical CC method (TMCM) with a human circulation hardware model based on hardware simulations. A benefit factor (BF), risk factor (RF) and benefit versus risk index (BRI) were introduced in this paper for the comprehensive evaluation of risk and benefit. The OCC-MCC was developed using the LabVIEW control platform and the mechanical chest compressor (MCC) controller. PID control is also employed by MCC for effective compression depth regulation. In addition, the physiological parameters model for MCC was built based on a digital signal processor for hardware simulations. A comparison between the OCC-MCC and TMCM was then performed based on the simulation test platform which is composed of the MCC, LabVIEW control platform, physiological parameters model for MCC and the manikin. Compared with the TMCM, the OCC-MCC obtained a better trade-off and a higher BRI in seven out of a total of nine cases. With a higher mean value of cardiac output (1.35 L/min) and partial pressure of end-tidal CO2 (15.7 mmHg), the OCC-MCC obtained a larger blood flow and higher BF than TMCM (5.19 vs. 3.41) in six out of a total of nine cases. Although it is relatively difficult to maintain a stable CC depth when the chest is stiff, the OCC-MCC is still superior to the TMCM for performing safe and effective CC during CPR. The OCC-MCC is superior to the TMCM in performing safe and effective CC during CPR and can be incorporated into the current version of mechanical CC devices for high quality CPR, in both in-hospital and out-of-hospital CPR settings.

[Application of the inspiratory impedance threshold device and its research progress].

The inspiratory impedance threshold device (ITD) was put forward by Lurie in 1995, and was assigned as a class II a recommendation by the International Liaison Committee on Resuscitation (ILCOR) resuscitation guidelines in 2005. The ITD is used to augment negative intrathoracic pressure during recoil of the chest so as to enhance venous return and cardiac output, and to decrease intracranial pressure. In the recent years many researches on the ITD have been1 carried out, but all the researches can not take out a clear evidence to support or refute the use of the ITD. This paper introduces the structure and working principle of the ITD in detail, the research results and the debates about the use of the ITD for the past years.

Buccal partial pressure of carbon dioxide outweighs traditional vital signs in predicting the severity of hemorrhagic shock in a rat model.

BACKGROUND: Hemorrhagic shock (HS) is a leading cause of death in both military and civilian settings. Researchers have investigated different parameters as predictors of HS, but reached inconsistent conclusions. We hypothesized that buccal partial pressure of carbon dioxide (PCO2) was a better predictor of HS than traditional vital signs. MATERIALS AND METHODS: Twenty-four anesthetized Wistar rats were randomly divided into four groups: one control group (no bleeding) and three surgical groups (25%, 35%, and 45% blood loss). Hemorrhage was induced by withdrawing blood from the left femoral artery over a period of 30 min. After that, resuscitation was performed on animals in surgical groups using the Ringer lactate solution. Buccal PCO2 was continuously measured by a newly designed sensor holder during the experiments. Traditional vital signs, cardiac output, base excess, and microvascular perfusion (MPF) were also measured or calculated. RESULTS: Buccal PCO2 differed significantly among four groups beginning at 20 min, approximately 10 min earlier than the shock index and more earlier than the heart rate, systolic blood pressure, and mean arterial pressure. Buccal PCO2 correlated well with cardiac index and the changes in MPF. The correlation coefficients with cardiac index, chest MPF, and upper-limb MPF for buccal PCO2 were 0.781, -0.879, and -0.946, respectively. Besides, buccal PCO2 showed a good value for predicting mortality. Furthermore, an approximate critical threshold of buccal PCO2 was also identified for predicting the severity of HS. CONCLUSIONS: Buccal PCO2 was a noninvasive, sensitive indicator of HS than traditional vital signs and may help on-scene rescuers administer early treatment of injured patients.

[Manual chest compression depth estimation based on integration reset mechanism].

To realize the measurement of the chest compression depth during the administration of manual cardiopulmonary resuscitation, two 3-axis digital accelerometers were applied for chest compression acceleration and environment acceleration acquisition, with one placed in the chest compression sensor pad, and the other placed in the back sensor pad. Then double integration was made for the acceleration-to-depth conversion with both of the accelerations after preprocessing. The method further included integration reset mechanism based on compression force, with the force point of a pre-determined threshold and the maximum force point as the starting point and the ending point of the integration, respectively. Moreover, a software compensation algorithm was implemented to further increase the accuracy of the depth estimation and reliability of the acceleration. The final performance of the compression depth estimation is within +/- 0.6 cm with 95% confidence of a total of 283 compressions. Accurate and real-time estimation of chest compression depth greatly facilitates the control of compression depth for the lifesaver during manual cardiopulmonary resuscitation.

[The development of a portable life support device for transporting pre-hospital critically ill patients].

OBJECTIVE: To describe a portable life support device for transportation of pre-hospital patients with critical illness. METHODS: The characteristics and requirements for urgent management during transportation of critically ill patients to a hospital were analyzed. With adoption of the original equipment, with the aid of staple of the art soft ware, the overall structure, its installation, fixation, freedom from interference, operational function were studied, and the whole system of life support and resuscitation was designed. RESULTS: The system was composed by different modules, including mechanical ventilation, transfusion, aspiration, critical care, oxygen supply and power supply parts. The system could be fastened quickly to a stretcher to form portable intensive care unit (ICU), and it could be carried by different size vehicles to provide nonstop treatment by using power supply of the vehicle, thus raising the efficiency of urgent care. CONCLUSION: With characteristics of its small size, lightweight and portable, the device is particularly suitable for narrow space and extreme environment.

[Development of portable free hemoglobin detector].

The paper designed the portable free hemoglobin detector using spectrophotometry, which determines the concentration of free hemoglobin. The device has features of portability, compact and precise. It can evaluate the quality of blood products instantly and effectively.

An optimal closed-loop control strategy for mechanical chest compression devices: a trade-off between the risk of chest injury and the benefit of enhanced blood flow.

OBJECTIVES: The widespread application of chest compression (CC) as a first aid measure inevitably has the potential for both harm and benefit. The present study was therefore undertaken to design an optimal CC closed-loop control strategy (OCCCS) for mechanical CC devices that will present an effective trade-off between the risk of chest injury and the benefit of blood flow during CPR. Additionally, to evaluate the CC performance of the OCCCS, the differences between the OCCCS and the traditional mechanical CC method (TMCM) of performing standard CC were explored. METHODS: Using the computer simulation technique, partial pressure of end-tidal CO2 (PETCO2) and human chest stiffness are simulated based on the Babbs' model in present study. PETCO2 was regarded as a benefit factor (BF), which was divided into 3 levels, while chest stiffness was regarded as a risk factor (RF), which was divided into 4 levels. A benefit versus risk index (BRI) was also constructed for the comprehensive evaluation of risk and benefit. An OCCCS was developed with the combination of the BF, RF, BRI and fuzzy control strategy. A comparison between the OCCCS and TMCM was then performed based on computer simulations. RESULTS: The OCCCS obtained a greater BRI and a better trade-off between risk and benefit than the TMCM in 6 out of a total 9 cases, and the OCCCS also resulted in a significantly improved cardiac output (CO) and PETCO2 in 6 of the 9 cases. The mean BRI, CO and PETCO2 resulting from the OCCCS were 5.69, 1.45 L/min and 15.51 mmHg, respectively, while the mean BRI, CO and PETCO2 resulting from TMCM were 4.76, 1.18 L/min and 13.26 mmHg, respectively. CONCLUSIONS: The OCCCS can provide safer and more effective CC during cardiopulmonary resuscitation (CPR) compared to the TMCM, and has great potential in the future mechanical CC device development.

[Computer simulation study on physiological feedback parameters during chest compression].

To have a thorough understanding of the CPR quality based on patients' various physiological states, the doctors must do something to simulate the chest compression physiological feedback parameters (CCPFP). The CCPFP simulation plays an important role in raising efficiency of CPR training and improving chest compression quality. In this study, the CCPFP, including cardiac output (CO), coronary perfusion pressure (CPP), partial pressure of End-tidal CO2 (PETCO2) and mean arterial relaxation pressure (MARP), was simulated using Charles F. Babbs' Model. Simulation results showed that the effect of compression depth upon CCPFP was important in the range of 2-6 cm, whereas compression rate had little effect on the CCPFP higher than 100/min; the thoracic factor is inversely proportional to the CCPFP with fixed compression depth and compression rate. The CCPFP simulation can be implemented at the various physiological statuses, and verified well with the animal experimental results and the clinical results.