A device for guided cpr

Abstract

The present invention relates to CPR measuring device (1) adapted to provide information to a person performing CPR, to a person operating a mechanical compression device or directly to a mechanical compression device. The CPR measuring device is adapted with a sensor (SENS), preferably a PPG sensor, a processor and a flexible sheet (F_S). The flexible sheet is shaped so as to be positioned on the chest of a person suffering from e.g. cardiac arrest, and wherein a user interface of the CPR measuring device is adapted to provide feedback with respect to the quality of performed CPR. The device is configured to measure a signal representing a difference in blood flow during chest compressions and provide a person with information relating to the efficiency or quality of the performed CPR.

Description

[0001] 85398PC01

[0002] 1

[0003] A DEVICE FOR GUIDED CPR

[0004] FIELD OF THE INVENTION

[0005] The present invention relates to cardiopulmonary resuscitation (CPR) during cardiac arrest. In particular, the invention relates to a device and method adapted to guide a person performing CPR, thus improving both short term and long term survival rates of cardiac arrest, by improving the quality of first responder CPR.

[0006] BACKGROUND OF THE INVENTION

[0007] Cardiopulmonary resuscitation (CPR) is a life-saving technique used in emergencies when someone's heartbeat has stopped. This can occur due to various reasons, such as a heart attack, drowning, or choking. The primary goal of CPR, to a first responder, is to maintain vital blood flow to the brain and other organs until professional medical help can arrive and restore normal heart function.

[0008] CPR works by combining chest compressions and rescue breaths. Chest compressions help to manually pump blood through the heart and maintain circulation, while rescue breaths provide oxygen to the lungs. The combination of these actions helps to keep the body's vital organs supplied with oxygenated blood, which is crucial for preventing brain damage and increasing the chances of survival.

[0009] The history of CPR dates back several centuries, with various techniques being developed and refined over time. In the 18th century, mouth-to-mouth resuscitation was recommended by the Paris Academy of Sciences for drowning victims.

[0010] CPR guidelines have continued to evolve over the years, with organizations such as the American Heart Association (AHA) and the American Red Cross playing key roles in updating and disseminating these guidelines. In 1963, the AHA endorsed CPR and launched a program to teach physicians about closed-chest cardiac resuscitation. 85398PC01

[0011] 2

[0012] By the early 1970s, CPR training was being offered to both laypeople and professionals, making it more widely accessible. In recent years, CPR guidelines have been updated several times, to emphasize the importance of high-quality chest compressions and the use of automated external defibrillators (AEDs). The focus has shifted towards minimizing interruptions in chest compressions and ensuring that compressions are performed at the correct depth and rate. Additionally, hands-only CPR, which involves chest compressions without rescue breaths, has been promoted for untrained bystanders to increase the likelihood of bystander intervention in cardiac emergencies.

[0013] Overall, the evolution of CPR guidelines reflects ongoing research and advancements in our understanding of resuscitation techniques. These updates aim to improve the effectiveness of CPR and increase the chances of survival for individuals experiencing cardiac arrest.

[0014] The above results in often changing guidelines, rendering previous training less efficient relative to current standards. Even further, performing chest compressions with a too high force may be harmful to the person receiving the CPR, whereas too low force or CPR performed at an incorrect location of the chest may not aid the restoration of blood flow, and may damage organs of the person receiving CPR, all of which may prevent a first responder, such as a bystander, from attempting to perform CPR.

[0015] OBJECT OF THE INVENTION

[0016] It is an object of the present invention to provide a device and method for guiding a first responder performing CPR.

[0017] SUMMARY OF THE INVENTION

[0018] The above-described object and several other objects are intended to be obtained in a first aspect of the invention by providing a cardiopulmonary resuscitation CPR measuring device, the CPR measuring device comprising:

[0019] -a sensor, the sensor adapted to generate a signal in response to a detected blood flow of a host, 85398PC01

[0020] 3

[0021] -a flexible sheet adapted to be positioned on the chest of the host, covering at least a part of said chest, the flexible sheet comprising one or more load cells,

[0022] -a processor in data connection with at least the sensor and the flexible sheet, the processor comprising a mathematical model adapted to convert the signal received from the sensor into a relative blood flow, correlating the relative blood flow to a signal received from the one or more load cells, and

[0023] -a user interface in data connection with the processor, the user interface adapted to provide a user, during cardiopulmonary resuscitation of the host, with an output, the output indicating whether an amplitude of the signal from the sensor is trending upwards, downwards or being stable.

[0024] The invention is particularly, but not exclusively, advantageous for obtaining better guidance for first responders, prior to an ambulance or other trained medical professionals arriving at the patient. The invention is further advantageous for removing a bystander's fear of performing CPR, as the present invention will assist said bystander.

[0025] Further, the present invention is particularly advantageous, in that it may reduce the risk of a first responder performing harmful CPR, by guiding both position and depth of the compressions performed during CPR.

[0026] Even further, the assistance provided by the present invention is particularly advantageous for improving the quality of CPR performed, by ensuring that the CPR is performed with an optimal force, i.e. to an optimal depth, at an optimal position on the chest and at an optimal rate.

[0027] Even further, the assistance provided by the present invention ensures that compressions during prolonged CPR are maintained at an optimal position, at an optimal pressure / depth and with an optimal frequency. It is to be understood, that performing CPR for a prolonged period of time is very straining, and thus, without assistance, may vary in quality, over time. 85398PC01

[0028] 4

[0029] In the context of the present invention, a load cell is to be understood as a sensor adapted to provide a signal when provided with a force, i.e. a compression, and wherein the load cell is adapted to convert the compressions into force. In some embodiments, the load cell or sensor provided a signal with an amplitude, and wherein the mathematical model is adapted to convert the signal into force.

[0030] Even further, it is to be understood, that the mathematical model may be adapted to convert force into depth of a compression, i.e. the depth that the hands move towards the thoracic cavity, in a ventral-dorsal direction. It is to be understood, that the mathematical model does not necessarily need the actual depth, as the force applied is sufficient for the device to function properly. Nevertheless, force may, in specific examples be correlated to depth, if the chest compression resistance is known.

[0031] In some embodiments, the load cells may be selected from one or more of strain gauge sensors, piezoelectric sensors, capacitive sensors, force-sensing resistors or magnetostrictive sensors.

[0032] In particular, strain gauge sensors may be advantageous, as they are reliable and cost-effective, which may enable the present invention to become more common, thus reducing the risk of fatal outcomes from e.g. cardiac arrest.

[0033] In particular, piezoelectric sensors may be advantageous, as they work well for sensing dynamic changes and dynamic pressure measurements.

[0034] In particular, capacitive sensors may be advantageous, as they provide a high sensitivity and precision to force applied.

[0035] In particular, force-sensing or force-sensitive resistors may be advantageous, as the flexible sheet may be made from a conductive polymer, and wherein the surface of the flexible sheet functions as load cells. This may provide a higher resolution than an array of other types of sensors, increasing the resolution of the signal provided by the conductive polymer. 85398PC01

[0036] 5

[0037] In particular, magnetostrictive sensors may be advantageous, as they provide hight durability and reliability.

[0038] In a preferred embodiment, the sensor is an on-skin sensor adapted to be releasably fixated to the skin of the host. This embodiment is particularly advantageous for ensuring a proper measurement of the blood flow. The on-skin sensor in combination with measurements provided by the one or more load cells.

[0039] In preferred embodiments, the flexible sheet comprises at least two load cells, such as a plurality of load cells.

[0040] In another preferred embodiment, the flexible sheet comprises 2x2 load cells arranged in an array, such as 3x3 load cells arranged in an array, such as 4x4 load cells arranged in an array, such as 5x5 load cells arranged in an array, such as 6x6 load cells arranged in an array, such as 2x3 load cells arranged in an array, such as 2x4 load cells arranged in an array, such as 3x4 load cells arranged in an array, such as 3x5 load cells arranged in an array, such as 3x6 load cells arranged in an array, such as 4x6 load cells arranged in an array, such as 4x7 load cells arranged in an array, such as 5x8 load cells arranged in an array, such as 7x8 load cells arranged in an array, such as 8x9 load cells arranged in an array. Preferably, the flexible sheet comprises between 4x6 and 5x8 load cells arranged in an array.

[0041] Preferably, the flexible sheet comprises between 30 and 100 load cells arranged in an array, more preferably the flexible sheet comprises between 40 and 80 load cells arranged in an array.

[0042] It is to be understood, that a higher number of rows and columns of the array represents a higher resolution, in respect to measuring an accurate position of the compressions on the chest.

[0043] It is further to be understood, that the flexible sheet may be utilized both on skin and on clothes of a patient. This may be particularly advantageous, in that a bystander may be conscious about providing CPR to a subject of the opposite sex, thus further increasing the likelihood of the first responder being a close bystander, reducing response time, i.e. time from detected cardiac arrest and the commencement of CPR. 85398PC01

[0044] 6

[0045] Even further, it is to be understood, that the flexible sheet may provide a visible and fluid barrier between the skin of the host and the first responder, which may further increase the likelihood of a close bystander commencing CPR.

[0046] It is further to be understood, that the flexible sheet may ensure that the host retains body heat during CPR, thus reducing the risk of hypothermia and further to reduce the risk of vasoconstriction.

[0047] In an advantageous embodiment, the flexible sheet comprises at least two layers. In some embodiments, the one or more load cells may be positioned in between said two layers. This embodiment is particularly advantageous for providing protection to the load cells, such as from sweat or other liquids, which may damage the load cells.

[0048] In other preferred embodiments, the flexible sheet comprises conductive polymers, adapting the flexible sheet as one or more force-sensing I force sensitive resistors. This embodiment may be particularly advantageous, as this may replace the one or more load cells.

[0049] In other advantageous embodiments, the processor and / or user interface may be integrated within the flexible sheet, thus ensuring a reduced complexity of the CPR measuring device, to a first responder, which may be anxious or stressed, due to the stressful situation of experiencing a person to suffer from cardiac arrest.

[0050] Thus, the present invention provides a simple and effective means for reducing the stress and anxiety of experiencing someone suffering from e.g. cardiac arrest, thus easing the stress and overall resulting in better decision making and calmness in a very stressful situation.

[0051] In the context of the present invention, relative blood flow is to be understood as a temporal difference in blood flow, i.e. wherein a signal representing a blood flow indicates that a compression has been performed. This, correlated with a signal from the load cell, is utilized to determine that the signal represents a true 85398PC01

[0052] 7 measurement of a blood flow different from zero. Further, the relative blood flow enables the device to determine the quality of the CPR, by measuring the difference in blood flow during CPR, as the difference in signal amplitude from the sensor provides a measure to the processor, which can be used to determine the effectiveness of a compression, i.e. with too low force there will be little variation in blood flow, with higher force there will be a higher variation in blood flow, with too high force there will be no further increase in blood flow variation.

[0053] In yet another preferred embodiment, the sensor is a photoplethysmogram PPG sensor. This embodiment is particularly advantageous, in that such sensor is cheap, durable, flexible, light weight and easy to apply to the skin of a host.

[0054] A PPG sensor optically obtains a plethysmogram. The PPG sensor works by illuminating the skin with light, usually from a light-emitting diode (LED), and then measuring the amount of light either transmitted or reflected to a photodetector. The changes in light absorption are caused by the varying blood volume in the tissue, which occurs with each heartbeat. PPG sensors are commonly used in devices like pulse oximeters to monitor heart rate and blood oxygen saturation. They are non-invasive, low-cost, and provide valuable information related to the cardiovascular system.

[0055] The inventors have realized, that by correlating a PPG signal with a true blood flow sensor, a mathematical model can be derived, the mathematical model enabling the PPG sensor to function as a blood flow sensor adapted to measure a relative blood flow. Using the PPG signal as an indicator of generated blood flow, the system can determine if changes in CPR location result in better or worse generated blood flow.

[0056] The effectiveness of CPR by means of the PPG derived flow signal can be determined by analysis of the sinusoidal frequency and phase content of local sections of the PPG signal over time, e.g. short-time Fourier transform.

[0057] Here the energy level of such frequency ranges that correspond to the applied CPR compressions are compared with the average energy level of the remaining frequency ranges, e.g. by calculating the ratio. 85398PC01

[0058] 8

[0059] Frequency ranges that relate to the CPR compression activity can be the actual frequency of compression and its higher harmonics. In tests performed by the inventors, the 1stand 2ndhigher harmonics appear very indicative.

[0060] The advantage of investigating the energy levels of frequency ranges instead of time-domain-based analysis, e.g. extraction of the peak-to-peak value, is that it is not distorted by short, artifact-like variations of the signal, that can happen regularly with every compression but do not significantly increase the mean value of the pulse curve and hence have little influence on flow.

[0061] It is to be understood, that a subject suffering from cardiac arrest presents with a blood flow at or near zero. Thus, the PPG can be used to provide a signal, as a response to a detected blood flow during a compression. As further compressions are applied, the mathematical model can correlate compare signal amplitudes from the PPG, to determine whether blood flow increases or decreases during said compressions. Even further, the mathematical model is adapted to correlate compressions measured by the one or more load cells with the signals provided by the PPG, based on the correlations, provide a first responder, which performs CPR, as to optimizing one or more of position of the compressions, depth of the compressions and frequency of the compressions.

[0062] In an advantageous embodiment, the sensor comprises an adhesive, the adhesive adapted to adhere to skin of the host. This embodiment is particularly advantageous for ensuring, that the sensor is sufficiently fixated to the skin to perform measurements. Even further, the adhesive enables a user to easily switch a position of the sensor, should the sensor be positioned in an area with low perfusion.

[0063] In another advantageous embodiment, the sensor comprises an adhesive patch with a foil, the sensor adapted to activate the device upon removal of the foil. This embodiment is particularly advantageous, in that the foil ensures that the sensor is positioned on the host, i.e. is not forgotten / f unctions as a fail safe. Further, the foil is advantageous in maintaining the adhesive properties of the adhesive patch until application on a host. 85398PC01

[0064] 9

[0065] In yet another advantageous embodiment, the flexible sheet further comprises a skin surface and a top surface, the top surface comprising an array of lights, such as LED's. This embodiment is particularly advantageous in providing an intuitive means of indicating to a first responder, where to optimally positioning the hands to perform compressions. It is to be understood, that the user interface may be construed as the array of lights.

[0066] In a preferred embodiment, the CPR measuring device further comprises an auditory device, such as a speaker. This embodiment is particularly advantageous to support the user interface, in providing assistance to the first responder during CPR. Even further, should the first responder not react to the assistance provided, e.g. due to stress or other factors, the auditory device may generate a sound to direct the attention of the first responder towards the user interface.

[0067] It is to be understood, that the user interface may comprise a screen, monitor or other suitable device. In some embodiments, the user interface is integrated into the flexible sheet.

[0068] In a preferred embodiment, the user interface is a display or touch screen.

[0069] In other preferred embodiments, the user interface is positioned at or near the chin of the host, ensuring a line of sight between the first responder and the user interface.

[0070] It is to be understood, that the flexible sheet may be uniform in shape.

[0071] In a preferred embodiment, the flexible sheet is shaped so as to clearly demonstrate how the flexible sheet is to be positioned on the thorax of the patient. This embodiment is particularly advantageous to ensure that CPR is performed quickly after retrieving the CPR measuring device.

[0072] It is further to be understood, that the combination of the sensor and the load cells, ensure, that the mathematical model can adapt to a compression location even if the flexible sheet is sliding from its original position, during CPR. 85398PC01

[0073] 10

[0074] In some embodiments, the CPR measuring device further comprises a second sensor in data connection with at least the processor, the second sensor adapted to be positioned at a distance from a first sensor, the processor adapted to compare measured signals received from the sensors, respectively. This embodiment may be particularly advantageous for providing a more accurate measurement of relative blood flow.

[0075] In preferred embodiments, the second sensor may be a sensor adapted to measure CO2, such as an expiration fraction of CO2, during CPR.

[0076] The second sensor may be selected from e.g. a near-infrared sensor or an impedance cardiography sensor.

[0077] In particular, an impedance cardiography sensor, as a patch applied to the neck or throat, near the carotid artery, may be advantageous to provide the processor with further data related to the quality of CPR.

[0078] In an advantageous embodiment, the CPR measuring device further comprises a data input / output device, such as a wireless receiver or wireless transceiver. It is to be understood, that the device may be configured to receive over-the-air updates, such as updated guidelines on CPR derived from future studies performed. This embodiment is particularly advantageous for ensuring, that a first responder, when using the present invention, always performs optimal CPR according to most recent guidelines, thus continuously increasing the quality of CPR.

[0079] In other advantageous embodiments, the transceiver is configured to receive information / , and adapted to send information / data. Such embodiments are particularly advantageous, in that data from the device, during or after performed CPR, may be subject to analysis, for further improvements.

[0080] Further, data from the device may be forwarded to e.g. an ambulance or other medical professionals enroute, to better prepare the medical professionals on the task at hand. 85398PC01

[0081] 11

[0082] In another advantageous embodiment, the user interface is adapted to receive information regarding the host, the information selected from one or more of: gender, age, height and weight. This embodiment may be particularly advantageous for optimizing the CPR. Furthermore, should future studies realize that CPR is to be specifically according to gender, the CPR measuring device may be adapted to receive over-the-air updated according to such studies.

[0083] In preferred embodiments, the mathematical model is a learning algorithm able to determine quality of the generated blood flow during CPR by analysing blood flow waveforms and classify them. In some embodiments, the classification comprises e.g. good or bad.

[0084] In a preferred embodiment, the mathematical model is a learning algorithm adapted to continuously optimize blood flow during CPR. It is to be understood, that an optimal blood flow during CPR is the product of a plurality of parameters, such as, but not limited to age, weight, gender and height. Furthermore, the quality of the CPR is a product of at least position of compressions, force / depth of the compressions and the frequency of the compressions. By applying a learning algorithm to the CPR measuring device enables the processor to adapt the assistance to specific parameters, based on the measurements performed, and tailor the CPR to the specific host.

[0085] Other advantages of the learning algorithm, in particular when functioning with a wireless data transceiver, is the ability of further CPR measuring devices to continuously provide improved mathematical models to faster and more effectively determine a best position of compressions, force of compressions and frequency of compressions, relative to different subjects suffering from a cardiac arrest, i.e. wherein data from performing CPR to subjects of different age, height, weight, gender are used between several devices to continuously optimize CPR in future similar medical emergencies.

[0086] It is to be understood, that the mathematical model or learning algorithm may be adapted to upload data to a central server or cloud solution, for further analysis and training of further learning algorithms. 85398PC01

[0087] 12

[0088] In preferred embodiments, the learning algorithm is a pattern recognition algorithm, trained on data from CPR, to provide optimal feedback to a person performing CPR.

[0089] In preferred embodiments, the learning algorithm is a neural network or convolutional neural network trained on data from CPR, to provide optimal feedback to a person performing CPR.

[0090] In preferred embodiments, the convolutional neural network is configured to perform one-dimensional (ID) convolution operations on blood flow waveforms, thereby extracting salient features which are subsequently analysed by downstream layers of the model.

[0091] In further preferred embodiments, the learning algorithm is adapted to recognise and classify patterns within the blood flow waveform produced by a chest compression, by comparing said waveform to previously learned reference waveforms corresponding to compressions identified as either effective (classified as "good") or ineffective, such as those arising from liver compression (classified as "bad").

[0092] In yet further preferred embodiments, training of the learning algorithm is conducted utilising a cross-entropy loss function, thereby enhancing the predictive accuracy of the model.

[0093] In preferred embodiments, the learning algorithm is trained on animals or humans suffering from cardiac arrest, and wherein a blood flow measuring device, such as an ultrasound transducer has been used to provide true blood flow, while correlating the data with the signal from the PPG sensor.

[0094] In preferred embodiments, the learning algorithm is trained on data from compressions performed on top of the heart and / or near the liver.

[0095] In alternative embodiments, other types of machine learning models may be employed to perform classification of blood flow waveforms. For example, recurrent neural networks (RNNs) may be utilized due to their capability to process time-series data by maintaining internal memory states that capture temporal dependencies. A specific subclass of RNNs, known as long short-term 85398PC01

[0096] 13 memory (LSTM) networks, may be employed to improve performance in scenarios involving longer or more complex waveform sequences. Both RNNs and LSTMs have demonstrated efficacy in time-series analysis and may be suitable for implementation depending on the complexity and duration of the waveform data.

[0097] In yet another embodiment, transformer-based models may be employed for time-series classification. These models are capable of learning long-range dependencies within the data and may be advantageous in identifying subtle or distributed patterns across waveform sequences. However, in scenarios where the available training data is limited, the benefits of transformer architectures may be constrained.

[0098] In yet another embodiment, simpler classification models may be utilized. For instance, each waveform may be represented as a feature vector, and a k-nearest neighbor (k-NN) algorithm may be applied to group similar waveforms into clusters. Alternatively, a support vector machine (SVM) may be trained to separate waveform data into distinct classes. Such models may be particularly effective in scenarios involving a limited number of waveform classes, such as distinguishing between waveforms generated by near-heart and near-liver compressions. However, the utility of these models may diminish as the number of waveform classes increases, for example, when data is collected from additional anatomical locations.

[0099] In other preferred embodiments, the CPR measuring device further comprises one or more of a battery and a memory, enabling the electronic parts of the CPR measuring device to function properly, and preferably wirelessly.

[0100] It is further to be understood, that the processor may be integrated in an electronic circuit board further comprising memory and other components necessary to execute the mathematical model and the programming adapted to input date from sensors and output assisting information to a first responder.

[0101] In a second aspect, the invention relates to a method of assisting a user performing CPR to a patient suffering from cardiac arrest, the method comprising: -applying the CPR measuring device according to the first aspect, to the patient, 85398PC01

[0102] 14

[0103] -receiving a first signal from the sensor and the one or more load cells, -correlating the first signals from the sensor and the one or more load cells to detect whether chest compressions, performed by the user to the patient, provides a blood flow to the patient, and

[0104] -providing the user with an output indicative of the detected or lack of detected blood flow of the patient.

[0105] It is to be understood, that the assistance to the user performing CPR is to be construed as decision support and thus, the second aspect may alternatively be phrased as: A method of providing decision support to a user performing CPR to a patient suffering from cardiac arrest, the method comprising:

[0106] -applying the CPR measuring device according to the first aspect, to the patient,

[0107] -receiving a first signal from the sensor and the one or more load cells, -correlating the first signals from the sensor and the one or more load cells to detect whether chest compressions, performed by the user to the patient, provides a blood flow to the patient, and

[0108] -providing the user with an output indicative of the detected or lack of detected blood flow of the patient

[0109] In a preferred embodiment, the method further comprises receiving a second signal from the sensor and the one or more load cells, then converting the first and second signal from the sensor into a relative blood flow by a comparison of an amplitude of said signals and providing the user with a measurement indicative of relative blood flow of the patient.

[0110] In another preferred embodiment, the method further comprises an indication to the user, that an increase or decrease of the force of a compression performed to the patient's chest can optimize blood flow.

[0111] In yet another preferred embodiment, the method further comprises an indication to the user, that a change of position of compressions performed to the patient's chest can optimize blood flow. 85398PC01

[0112] 15

[0113] In an advantageous embodiment, the method further comprises an indication to the user, that an increase or decrease of frequency of compressions performed to the patient's chest can optimize blood flow.

[0114] The above indications to the user, based on relative blood flow measurements, are particularly advantageous in continuously improving the quality of CPR.

[0115] In another advantageous embodiment, the method further comprises an indication to the user, that a change of position of compressions after 10 to 20 seconds, from receiving the first signal from the sensor and the one or more load cells, can provide the processor with an optimal data set for optimizing blood flow. It is to be understood, that the indication may be adapted to reflect future guidelines on compressions per ventilation, in order to not disturb the first responder during a sequence of compressions.

[0116] In yet another advantageous embodiment, the method further comprises an indication of a second position of compressions to the user, the second position being different from a current position of compressions, the indication performed by activating one or more of the lights of the flexible sheet.

[0117] The above embodiments are particularly advantageous for ensuring, that a correct position on the chest on which to perform compressions has been located. In particular, a distressed first responder may initiate compressions at a location not centred above the heart, and wherein the mathematical model, by receiving comparable signals from at least two positions, is enabled to determine an optimal position, depth and frequency of said compressions.

[0118] In a preferred embodiment, the method further comprises an indication to the user, after a specific amount of measured compressions, that ventilation to the patient is advised, according to national health guidelines. This embodiment may be particularly advantageous during prolonged application of compressions, as the blood and lungs will be depleted from oxygen over time, if not ventilated.

[0119] In a third aspect the invention relates to a kit of parts, the kit of parts comprising an automated external defibrillator (AED) and the device according to the first 85398PC01

[0120] 16 aspect of the invention. This embodiment is particularly advantageous, in that the CPR measuring device may increase the quality of CPR, which in some cases of cardiac arrest, is a prerequisite for applying an automated external defibrillator.

[0121] It is to be understood, that an AED will not apply a shock, if no electric activity (asystolic) can be detected by the AED. Furthermore, an AED will not shock a heart which has an electrical activity, that is, not ventricular fibrillation or ventricular tachycardia. Thus, by performing high quality CPR, the heart may regain the function required for the AED to perform a shock, if deemed necessary by the internal computer of the AED.

[0122] In a fourth aspect, the invention relates to use of the device, method or kit of parts according to any of the preceding aspects for quality assessment of CPR. It is to be understood, that the device may be particularly advantageous for training purposes by providing the processor with a simulated signal representing the signals provided by the sensor, thus providing students with comprehensible and direct feedback during training.

[0123] In a fifth aspect, the invention relates to treatment of cardiac arrest by using the device, method or kit of parts, according to any of the first, second and third aspect.

[0124] In a sixth aspect, the invention relates to a decision support system, the system adapted to provide decision support to a person performing CPR, by use of the system according to the first aspect.

[0125] In alternative embodiments, the device and method are adapted to provide feedback to a mechanical chest compression device or a health care professional operating a mechanical chest compression device.

[0126] The first, second, third, fourth, fifth and sixth aspect of the present invention may each be combined with any of the other aspects. These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter. 85398PC01

[0127] 17

[0128] BRIEF DESCRIPTION OF THE FIGURES

[0129] The device and method according to the invention will now be described in more detail with regard to the accompanying figures. The figures show one way of implementing the present invention and is not to be construed as being limiting to other possible embodiments falling within the scope of the attached claim set.

[0130] FIG. 1 shows the device positioned on a host, according to an embodiment of the invention;

[0131] FIG. 2 shows the device positioned on a host during compressions at a location on the chest, according to an embodiment of the invention;

[0132] FIG. 3 shows the device positioned on a host during compressions at another location on the chest, according to an embodiment of the invention;

[0133] FIG. 4 shows a schematic illustration of the device, according to an embodiment of the invention;

[0134] FIG. 5 is a flow-chart of a method according to the invention;

[0135] FIG. 6 shows a graph, according to a study performed by the inventors;

[0136] FIG. 7 shows another graph, according to the study performed by the inventors;

[0137] FIG. 8A shows the device positioned on a host, with the flexible sheet in one placement on the chest;

[0138] FIG. 8B shows the device positioned on a host, with the flexible sheet in another placement than in FIG. 8A on the chest;

[0139] FIG. 9 shows an example of blood flow waveforms used for training a learning algorithm model, with test data being near-heart compressions;

[0140] FIG. 10 shows an example of blood flow waveforms used for training a learning algorithm model, with test data being near-liver compressions; and

[0141] FIG. 11 shows an example of blood flow waveforms, where the compression area is unknown.

[0142] DETAILED DESCRIPTION OF AN EMBODIMENT

[0143] FIG. 1 shows the device 1 positioned on a host H, according to an embodiment of the invention. The flexible sheet F_S of the device 1 is positioned on the chest of the host H. The flexible sheet F_S has an array of load cells L_C arranged, spaced apart. The flexible sheet F_S furthermore has a user interface UI, at or near the 85398PC01

[0144] 18 chin of the host H. It is to be understood, that positioning the user interface UI, close to the face of the host H, enables a clear line of sight between the user interface UI and a person providing CPR (see FIG. 2 and FIG. 3), and further enables the person to monitor any changes in facial expression of the host H. The flexible sheet F_S furthermore has an array of lights LED arranged on the surface. It is to be understood, that the dense array of load cells L_C and lights LED enable for a high resolution of both force measurements to the processor (see FIG. 4) and indication of compression to the person providing the compressions. It is to be understood, that compressions are to be construed as one element of CPR; and wherein ventilation is the other element. In some embodiments, the processor may be integrated within the user interface UI. A sensor SENS is fixated to an ear or earlobe of the host H, to measure the pulse wave as proxy for blood flow. It is to be understood, that the earlobe is typically well perfused, and thus may be a suitable position for positioning of the sensor SENS, but other positions may be suitable as well. In other embodiments, the user interface UI of FIG. 1 is merely an electronics circuit assembly, and wherein the interface conferring information to the person providing CPR is the lights LED. Thus, in some embodiments, the user interface is the lights LED.

[0145] FIG. 2 shows the device 1 positioned on a host H during compressions at a location on the chest, according to an embodiment of the invention. The flexible sheet F_S of the device 1 is positioned on the chest of the host H. The flexible sheet F_S has an array of load cells L_C arranged, spaced apart. The flexible sheet F_S furthermore has a user interface UI, at or near the chin of the host H. The flexible sheet F_S furthermore has an array of lights LED arranged on the surface. A sensor SENS is fixated to an ear or earlobe of the host H, to measure blood flow. FIG. 2 illustrates a less than optimal compression position C_P, as the hands are not located above the heart, according to typical human anatomy. This is further to be seen, when comparing the graph of the user interface UI of FIG. 2, with the graph of the user interface UI of FIG. 3, as the amplitude of the graph of FIG. 3 is higher, displaying a larger variance in signal, compared to the graph of FIG. 2. It is to be understood, that the graph of the user interface UI shows signal amplitude or blood flow on the y-axis, and time on the x-axis. 85398PC01

[0146] 19

[0147] FIG. 3 shows the device 1 positioned on a host H during compressions at another location on the chest, according to an embodiment of the invention. The flexible sheet F_S of the device 1 is positioned on the chest of the host H. The flexible sheet F_S has an array of load cells L_C arranged, spaced apart. The flexible sheet F_S furthermore has a user interface UI, at or near the chin of the host H. The flexible sheet F_S furthermore has an array of lights LED arranged on the surface. A sensor SENS is fixated to an ear or earlobe of the host H, to measure pulse wave as proxy for blood flow.

[0148] The device 1 further comprises a processor PROC which, in preferred embodiments, is configured to interrogate the array of load cells L_C arranged within the flexible sheet F_S. Upon acquisition of measurement data from the load cells L_C, the processor PROC is operable to compute the centroid of the applied forces, thereby establishing the hand placement of a first responder administering chest compressions to the host H. Furthermore, the processor PROC is adapted to aggregate the force measurements in order to determine the magnitude of each chest compression imparted. The device 1 is additionally configured to assess the quality of cardiopulmonary resuscitation (CPR) by reference to pulse wave data as proxy for blood flow data acquired via the sensor SENS, which is preferably constituted as a photoplethysmography (PPG) sensor equipped with a PPG probe affixed to the host H prior to initiation of CPR. During the administration of CPR, each chest compression is measured by the device 1 and assigned a Quality-of- CPR score predicated upon the amplitude of the PPG signal (wherein a higher amplitude is indicative of improved quality) and the calculated compression force (wherein a higher force is considered less desirable). The system is further adapted to correlate the Quality-of-CPR scores to the specific anatomical locations on the chest of the host H.

[0149] FIG. 3 illustrates a near optimal compression position C_P, as the hands are located above the heart, according to typical human anatomy. This is further to be seen, when comparing the graph of the user interface UI of FIG. 3, with the graph of the user interface UI of FIG. 2, as the amplitude of the graph of FIG. 3 is higher, displaying a larger variance in signal, compared to the graph of FIG. 2.

[0150] FIG. 4 shows a schematic illustration of the device 1, according to an embodiment of the invention. The processor PROC is arranged to receive signals from the one or more load cells L_C and the sensor SENS. The processor PROC is further in 85398PC01

[0151] 20 data connection with a transceiver TRAN, which enables the processor PROC to communicate with a server SERV. The processor is furthermore adapted to provide an output to a person performing CPR through the user interface UI. In this particular embodiment, the processor PROC, the one or more load cells L_C, the transceiver TRAN and the user interface UI is integrated within the flexible sheet F_S, but other embodiments are to be construed, such as wherein the user interface UI is an auxiliary device.

[0152] FIG. 5 is a flow-chart of a method according to the invention. FIG. 5 shows a flowchart of a method of assisting a user performing CPR to a patient suffering from cardiac arrest, the method comprising the following steps:

[0153] 51-applying the CPR measuring device according to the first aspect to the patient,

[0154] 52-receiving a first signal from the sensor and the one or more load cells,

[0155] 53-correlating the first signals from the sensor and the one or more load cells to detect whether chest compressions, performed by the user to the patient, provides a blood flow to the patient, and

[0156] 54-providing the user with an output indicative of the detected or lack of detected blood flow of the patient.

[0157] In some embodiments, the system can be configured to display to the user a heat map indicating the locations on the mat where chest compressions should be performed. This is achieved by means of the integrated LEDs within the flexible sheet F_S.

[0158] FIG. 6 shows a graph, according to a study performed by the inventors. The graph shows the waveforms from simultaneously measured PPG-signal and blood flow, as well as an average level for the blood flow amount. The graph shows the recorded data from five chest-compressions administered approximately 10 centimetres closer to the abdomen on a pig with cardiac arrest, i.e. at a less-than optimal position for performing compressions, relative to the graph of FIG. 7.

[0159] The device 1 is further configured to non-invasively estimate changes in the blood flow generated during cardiopulmonary resuscitation (CPR) by utilising real-time values acquired from one or more photoplethysmography (PPG) sensors SENS while at the same time measuring actual blood flow by a flow transducer. There 85398PC01

[0160] 21 exists a positive correlation between the measured blood flow and the corresponding PPG-value, such that an increased pea k-to- peak amplitude in the PPG waveform is indicative of a greater generated blood flow and vice versa. A high measured blood flow and corresponding high PPG-value is an identification of a high degree of unidirectional flow.

[0161] FIG. 7 shows another graph, according to a study performed by the inventors. The graph shows the waveforms from simultaneously measured PPG-signal and blood flow, as well as an average level for the blood flow amount. The graph shows the recorded data from five chest-compressions administered directly on top of the heart and, i.e. at a more ideal position, relative to FIG. 6, as the blood flow is seen to be much greater, based on the waveforms. Additionally, the PPG- waveform has a bigger amplitude which demonstrates a relationship between PPG-signal and generated blood flow.

[0162] FIG. 8A shows the device 1 positioned on a host H, with the flexible sheet F_S in one placement on the chest.

[0163] FIG. 8B shows the device 1 positioned on a host H, with the flexible sheet F_S in another placement than in FIG. 8A on the chest.

[0164] In certain embodiments, the flexible sheet F_S is secured to the host H by means of an adhesive applied to the reverse side thereof, thereby ensuring that the spatial arrangement of the array of load cells L_C and lights LED remains fixed relative to the host's chest throughout the resuscitation procedure. The device 1 does not depend upon any predetermined assumption regarding the optimal location for chest compressions, as anatomical variations may exist between different hosts H. Instead, device 1 is configured to identify the optimal compression site based on sensor data received from the sensor SENS and load cells L_C, and to employ integrated visual indicators, specifically the array of lights LED, to direct the first responder (i.e., the individual administering CPR) towards the optimal compression location. This configuration facilitates rapid application of the flexible sheet F_S, wherein the effectiveness of deployment is not contingent on the initial position or orientation of the sheet on the host's chest, provided that the flexible sheet F_S covers the as yet undetermined optimal chest location. 85398PC01

[0165] 22

[0166] In FIG. 8A, a first placement of the flexible sheet F_S upon the host H is depicted, whilst in FIG. 8B, the flexible sheet F_S is illustrated in an alternative position and orientation upon the host's H chest. In both embodiments, as demonstrated, the integrated visual indicators, namely the array of lights LED, remain consistent, thereby evidencing that the functionality of the device 1 is independent of the initial placement or orientation of the flexible sheet F_S.

[0167] FIG. 9 shows an example of blood flow waveforms used for training a learning algorithm model, with test data being near-heart compressions.

[0168] FIG. 10 shows an example of blood flow waveforms used for training a learning algorithm model, with test data being near-liver compressions.

[0169] FIG. 11 shows an example of blood flow waveforms, where the compression area is unknown.

[0170] FIG. 9, FIG. 10, and FIG. 11 respectively illustrate near-heart waveforms, near- liver waveforms, and waveforms of unknown origin, each derived from an animal trial utilising a human-sized pig model. In each figure, every individual waveform recorded during the trial is presented, with the specific intention of highlighting the variability inherent in these physiological responses. The average waveform can be seen in the figures indicated with the bold highlighted curve.

[0171] To ensure that the analysis conducted by the learning algorithm remains unbiased by differences in amplitude, which can be substantial between different patients, all waveforms have been normalised such that their peak values are set to 1. This normalisation process mitigates amplitude bias, allowing the algorithm to focus on the shape and pattern of the waveform rather than the absolute magnitude of the generated blood flow.

[0172] It is also important to note that the level of generated blood flow is subject to significant change, particularly as a patient's vascular system transitions into a low-flow state following cardiac arrest. This physiological alteration underscores the necessity of normalising the data, as it further minimises confounding effects arising from patient-to-patient variability in blood flow dynamics.

[0173] The purpose of the machine learning algorithm is to determine the quality of blood-flow generated during CPR by analyzing each waveform and identifying the 85398PC01

[0174] 23 presence of a dicrotic notch, a secondary peak caused by the closing of heart valves. This notch serves as an indicator of high-quality CPR, signifying unidirectional blood circulation through the vascular system. In contrast, waveforms lacking a dicrotic notch typically result from compressions of blood- filled organs such as the liver. Although such compressions may produce high PPG values, they do not reflect effective CPR, as the blood tends to move back and forth rather than being properly oxygenated in the lungs.

[0175] The example of FIG. 9-11 is one embodiment. In this embodiment, a dataset comprising 357 blood flow waveforms is utilized for training and evaluation of a machine learning model. The dataset includes three distinct subsets of waveforms: a first subset comprising 121 samples recorded during chest compressions performed near the heart (FIG. 9); a second subset comprising 174 samples recorded during compressions performed near the liver (FIG. 10), and a third subset comprising 62 samples also recorded near the heart (FIG. 11).

[0176] Two training configurations are employed. In a first configuration (Training Set A), all 357 waveforms are included. A portion of the data, corresponding to 20% of each subset, is randomly selected and reserved for testing, while the remaining 80% is used for training the model.

[0177] In a second configuration (Training Set B), only the first and second subsets (i.e., near-heart and near-liver compressions) are used for training, while the third subset is exclusively used for testing.

[0178] The learning algorithm is implemented as a convolutional neural network, configured to perform one-dimensional (ID) convolution operations on blood flow waveforms. These operations extract salient features that are analyzed by downstream layers to classify the waveform as either "good" (e.g., near-heart compressions) or "bad" (e.g., near-liver compressions), based on previously learned reference patterns. The CNN is trained using a cross-entropy loss function to enhance predictive accuracy.

[0179] The notch in the waveform at approximately 0.3 s is an identification of the degree of unidirectional flow. In FIG. 9 (near-heart compression), this notch is more prominent compared to FIG. 10 (near-liver compression), indicating a higher degree of unidirectional flow for the CPR given near-heart. 85398PC01

[0180] 24

[0181] Two versions of the CNN model are evaluated:

[0182] Model A is trained using Training Set A to assess the model's ability to distinguish between waveform types. The model demonstrates high classification accuracy, achieving approximately 96% accuracy on the test set. No significant overfitting is observed. A confusion matrix illustrating the classification results is shown below:

[0183] Accuracy=0.967

[0184] Precision = 1.00 Recall = 0.944

[0185] Fl score=0.971

[0186] Model B is trained using Training Set B to evaluate the model's generalization capability on previously unseen data. The model successfully classifies 61 out of 62 test waveforms as originating from near-heart compressions, indicating robust generalization performance.

[0187] 85398PC01

[0188] 25

[0189] The following is an itemized list of embodiments, Item 1 to Item 22, according to the invention.

[0190] Item 1. A cardiopulmonary resuscitation CPR measuring device, the CPR measuring device comprising:

[0191] -a sensor, the sensor adapted to generate a signal in response to a detected blood flow of a host,

[0192] -a flexible sheet adapted to be positioned on the chest of the host, covering at least a part of said chest, the flexible sheet comprising one or more load cells,

[0193] -a processor in data connection with at least the sensor and the flexible sheet, the processor comprising a mathematical model adapted to convert the signal received from the sensor into a relative blood flow, correlating the relative blood flow to a signal received from the one or more load cells, and

[0194] -a user interface in data connection with the processor, the user interface adapted to provide a user, during cardiopulmonary resuscitation of the host, with an output, the output indicating whether an amplitude of the signal from the sensor is trending upwards, downwards or being stable.

[0195] Item 2. The CPR measuring device according to Item 1, wherein the sensor is an on-skin sensor adapted to be releasably fixated to the skin of the host.

[0196] Item 3. The CPR measuring device according to any of Items 1 or 2, wherein the sensor is a photoplethysmogram PPG sensor.

[0197] Item 4. The CPR measuring device according to any of Items 1 to 3, the sensor comprising an adhesive, the adhesive adapted to adhere to skin of the host.

[0198] Item 5. The CPR measuring device according to any of the preceding Items, the sensor comprising an adhesive patch with a foil, the sensor adapted to activate the device upon removal of the foil. 85398PC01

[0199] 26

[0200] Item 6. The CPR measuring device according to any of the preceding Items, the flexible sheet further comprising a skin surface and a top surface, the top surface comprising the user interface, such as an array of lights, such as LED's.

[0201] Item 7. The CPR measuring device according to any of the preceding Items further comprising an auditory device, such as a speaker.

[0202] Item 8. The CPR measuring device according to any of the preceding Items further comprising a second sensor in data connection with at least the processor, the second sensor adapted to be positioned at a distance from the sensor according to Item 1, the processor adapted to compare measured signals received from the sensors, respectively.

[0203] Item 9. The CPR measuring device according to any of the preceding Items further comprising a data input / output device, such as a wireless receiver or wireless transceiver.

[0204] Item 10. The CPR measuring device according to any of the preceding Items, the user interface adapted to receive information regarding the host, the information selected from one or more of: gender, age, height and weight.

[0205] Item 11. The CPR measuring device according to any of the preceding Items wherein the mathematical model is a learning algorithm adapted to continuously optimize blood flow during CPR.

[0206] Item 12. A method of assisting a user performing CPR to a patient suffering from cardiac arrest, the method comprising:

[0207] -applying the CPR measuring device according to any of Items 1 to 11 to the patient,

[0208] -receiving a first signal from the sensor and the one or more load cells, -correlating the first signals from the sensor and the one or more load cells to detect whether chest compressions, performed by the user to the patient, provides a blood flow to the patient, and 85398PC01

[0209] 27

[0210] -providing the user with an output indicative of the detected or lack of detected blood flow of the patient.

[0211] Item 13. The method according to Item 12 further comprising:

[0212] -receiving a second signal from the sensor and the one or more load cells, -converting the first and second signal from the sensor into a relative blood flow by a comparison of an amplitude of said signals,

[0213] -providing the user with a measurement indicative of relative blood flow of the patient.

[0214] Item 14. The method according to Item 12 or 13 further comprising: -indicating to the user, that an increase or decrease of the force of a compression performed to the patient's chest can optimize blood flow.

[0215] Item 15. The method according to any of Items 12 to 14 further comprising: -indicating to the user, that a change of position of compressions performed to the patient's chest can optimize blood flow.

[0216] Item 16. The method according to any of Items 12 to 15 further comprising: -indicating to the user, that an increase or decrease of frequency of compressions performed to the patient's chest can optimize blood flow.

[0217] Item 17. The method according to any of Items 12 to 16 further comprising: -indicating to the user, that a change of position of compressions after 10 to 20 seconds, from receiving the first signal from the sensor and the one or more load cells, can provide the processor with an optimal data set for optimizing blood flow.

[0218] Item 18. The method according to any of Items 12 to 17 further comprising: -indicating a second position of compressions to the user, the second position being different from a current position of compressions, the indication performed by activating one or more of the lights of the flexible sheet. 85398PC01

[0219] 28

[0220] Item 19. The method according to any of Items 12 to 18 to further comprising:

[0221] -indicating to the user, after a specific amount of measured compressions, that ventilation to the patient is advised, according to national health guidelines.

[0222] Item 20. A kit of parts, the kit of parts comprising an automated external defibrillator and the device according to any of Items 1 to 11.

[0223] Item 21. Use of the device, method or kit of parts according to any of the preceding Items for quality assessment of CPR.

[0224] Item 22. Treatment of cardiac arrest by using the device, method or kit of parts, according to any of Items 1 to 20.

[0225] In short, the present invention relates to CPR measuring device 1 adapted to provide information to a person performing CPR, to a person operating a mechanical compression device or directly to a mechanical compression device. The CPR measuring device is adapted with a sensor SENS, preferably a PPG sensor, a processor and a flexible sheet F_S. The flexible sheet is shaped so as to be positioned on the chest of a person suffering from e.g. cardiac arrest, and wherein a user interface of the CPR measuring device is adapted to provide feedback with respect to the quality of performed CPR. The device is configured to measure a signal representing a difference in blood flow during chest compressions and provide a person with information relating to the efficiency or quality of the performed CPR.

[0226] Although the present invention has been described in connection with the specified embodiments, it should not be construed as being in any way limited to the presented examples. The scope of the present invention is set out by the accompanying claim set. In the context of the claims, the terms "comprising" or "comprises" do not exclude other possible elements or steps. Also, the mentioning of references such as "a" or "an" etc. should not be construed as excluding a plurality. The use of reference signs in the claims with respect to elements indicated in the figures shall also not be construed as limiting the scope of the 85398PC01

[0227] 29 invention. Furthermore, individual features mentioned in different claims, may possibly be advantageously combined, and the mentioning of these features in different claims does not exclude that a combination of features is not possible and advantageous.

Claims

85398PC0130CLAIMS1. A cardiopulmonary resuscitation CPR measuring device, the CPR measuring device comprising:-a sensor, the sensor adapted to generate a signal in response to a detected blood flow of a host,-a flexible sheet adapted to be positioned on the chest of the host, covering at least a part of said chest, the flexible sheet comprising one or more load cells,-a processor in data connection with at least the sensor and the flexible sheet, the processor comprising a mathematical model adapted to convert the signal received from the sensor into a relative blood flow, correlating the relative blood flow to a signal received from the one or more load cells, and-a user interface in data connection with the processor, the user interface adapted to provide a user, during cardiopulmonary resuscitation of the host, with an output, the output indicating whether an amplitude of the signal from the sensor is trending upwards, downwards or being stable.

2. The CPR measuring device according to claim 1, wherein the sensor is an on-skin sensor adapted to be releasably fixated to the skin of the host.

3. The CPR measuring device according to any of claims 1 or 2, wherein the sensor is a photoplethysmogram PPG sensor.

4. The CPR measuring device according to any of claims 1 to 3, the sensor comprising an adhesive, the adhesive adapted to adhere to skin of the host.

5. The CPR measuring device according to any of the preceding claims, the sensor comprising an adhesive patch with a foil, the sensor adapted to activate the device upon removal of the foil.

6. The CPR measuring device according to any of the preceding claims, the flexible sheet further comprising a skin surface and a top surface, the top85398PC0131 surface comprising the user interface, such as an array of lights, such as LED's.

7. The CPR measuring device according to any of the preceding claims further comprising an auditory device, such as a speaker.

8. The CPR measuring device according to any of the preceding claims further comprising a second sensor in data connection with at least the processor, the second sensor adapted to be positioned at a distance from the sensor according to claim 1, the processor adapted to compare measured signals received from the sensors, respectively.

9. The CPR measuring device according to any of the preceding claims further comprising a data input / output device, such as a wireless receiver or wireless transceiver.

10. The CPR measuring device according to any of the preceding claims wherein the mathematical model is a learning algorithm adapted to continuously optimize blood flow during CPR.

11. A method of assisting a user performing CPR to a patient suffering from cardiac arrest, the method comprising:-applying the CPR measuring device according to any of claims 1 to 10 to the patient,-receiving a first signal from the sensor and the one or more load cells, -correlating the first signals from the sensor and the one or more load cells to detect whether chest compressions, performed by the user to the patient, provides a blood flow to the patient, and-providing the user with an output indicative of the detected or lack of detected blood flow of the patient.

12. The method according to claim 11 further comprising:-receiving a second signal from the sensor and the one or more load cells, -converting the first and second signal from the sensor into a relative blood flow by a comparison of an amplitude of said signals,85398PC0132-providing the user with a measurement indicative of relative blood flow of the patient.

13. The method according to claim 11 or 12 further comprising: -indicating to the user, that an increase or decrease of the force of a compression performed to the patient's chest can optimize blood flow.

14. The method according to any of claims 11 to 13 further comprising: -indicating to the user, that a change of position of compressions performed to the patient's chest can optimize blood flow.

15. The method according to any of claims 11 to 14 further comprising: -indicating to the user, that a change of position of compressions after 10 to 20 seconds, from receiving the first signal from the sensor and the one or more load cells, can provide the processor with an optimal data set for optimizing blood flow.

Images