WO2017072626A1

WO2017072626A1 - System and method to automatically adjust, notify or stop chest compressions when a spontaneous pulse is detected during automated cardiopulmonary resuscitation

Info

Publication number: WO2017072626A1
Application number: PCT/IB2016/056248
Authority: WO
Inventors: Jens MÜHLSTEFF; Ralph Wilhelm Christianus Gemma Rosa WIJSHOFF; Eefje Janet Hornix
Original assignee: Koninklijke Philips NV
Current assignee: Koninklijke Philips NV
Priority date: 2015-10-27
Filing date: 2016-10-18
Publication date: 2017-05-04
Anticipated expiration: 2018-04-27

Abstract

A compression robot (10) automatically delivers cardiopulmonary resuscitation (CPR) chest compressions to a person. A physiological sensor (20, 30) measures a physiological signal (22, 32) of the person. An electronic processor (40, 70) is programmed to derive a spontaneous pulse representation from the physiological signal and control the compression robot based on the spontaneous pulse representation. A pulse rate of the spontaneous pulse representation may be determined, and a compression rate of the compression robot adjusted to a new compression rate that differs from the determined pulse rate by at least 10%. The electronic processor may be programmed to control the compression robot to stop the automatic delivery of CPR chest compressions to the person if the spontaneous pulse representation satisfies a return of spontaneous circulation (ROSC) criterion, or prompt to the caregiver that a potential ROSC has been detected so the caregiver can determine whether to stop or continue chest compressions.

Description

SYSTEM AND METHOD TO AUTOMATICALLY ADJUST, NOTIFY OR STOP CHEST COMPRESSIONS WHEN A SPONTANEOUS PULSE IS DETECTED DURING AUTOMATED CARDIOPULMONARY RESUSCITATION

FIELD

The following relates generally to the medical arts, medical emergency response arts, medical monitoring arts, cardiopulmonary resuscitation arts, and related arts.

BACKGROUND

Cardiopulmonary resuscitation (CPR) is indicated for cardiac arrest and is applied by manual chest compressions and rescue breaths. Current CPR guidelines call for alternating between 30 chest compressions followed by two ventilating breaths for a period of 2 min. After such a 2 min block, the electrocardiography (ECG) signal is analyzed, followed by a pulse check by manual palpation if an organized ECG rhythm has been observed. The goal of CPR is to achieve return of spontaneous circulation (ROSC) in which the beating heart drives life- sustaining blood circulation.

The quality of manual CPR chest compressions can be adversely impacted by lack of experience or skill of the emergency responder, or by the increasing exhaustion of the emergency responder if ROSC is not achieved in the first few minutes. These factors can lead to a non-uniform or too slow compression rate, and/or to shallow compressions that may be insufficient to drive life-sustaining blood circulation.

An automated CPR device comprises a compression robot strapped to the torso of the person in cardiac arrest. The compression robot includes a piston or other actuator that delivers CPR chest compressions at a precise compression rate and with precise depth. The compression robot does not tire, and can deliver high-quality CPR chest compressions essentially indefinitely. The automated CPR device also frees the emergency responder to perform other tasks.

If CPR achieves the goal of ROSC, the chest compressions should be stopped. Best practice recommendations vary: some recommend stopping chest compressions immediately upon detection of ROSC, while others recommend continuing compressions for up to a few tens of seconds after detection of ROSC to shortly extend the support to the recovering heart. Problems which can arise if CPR chest compressions are continued after ROSC is achieved include compression-induced cardiac re-fibrillation, or physical injury of the reviving person. Thus, it is important to quickly and accurately detect when ROSC is achieved by administered CPR and to stop chest compressions immediately or shortly after ROSC.

Detection of ROSC is typically done by manual palpation to detect spontaneous arterial blood pressure pulsations. The carotid artery, the femoral artery or the radial artery in the wrist is usually palpated for this purpose. If an automated external defibrillator (AED) is connected to the unconscious person receiving first aid (which is often the case as electric shock using the AED is the appropriate response if it is determined the unconscious person is in ventricular fibrillation) then the AED will acquire ECG information. However, the ECG is intended to detect a shockable cardiac rhythm. The ECG generally does not provide reliable detection of ROSC, because cardiac electrical activity, even if organized, does not necessarily translate into physical heart contractions effective to deliver life-sustaining circulation. By contrast, manual palpation directly detects spontaneous arterial blood pulsations. However, pulse check via manual palpation cannot be performed during delivery of CPR compressions because the emergency responder generally cannot distinguish arterial blood pulsations due to a spontaneous pulse from arterial blood pulsations produced by the CPR compressions.

In automated CPR, the emergency responder must therefore synchronize manual palpation pulse checks with pre-programmed periodic interruptions of the chest compressions delivered by the compression robot. Compared with manual CPR, this can actually increase the likelihood of failing to quickly detect ROSC, because the emergency responder may become distracted and fail to perform a pulse check during a pre-programmed compression interruption, leading to longer time intervals between pulse checks. More generally, automated compressions provide greater opportunity for the emergency responder to become distracted and fail to diligently monitor the arrested person for ROSC.

Accordingly, there remains a continued need for improved automated CPR techniques. BRIEF SUMMARY

In accordance with one illustrative example, a cardiopulmonary resuscitation (CPR) device comprises: a compression robot configured to automatically deliver CPR chest compressions to a person; a physiological sensor configured to measure a physiological signal of the person; and an electronic processor programmed to (i) derive a spontaneous pulse representation from the physiological signal and (ii) control the compression robot based on the spontaneous pulse representation. The electronic processor may be programmed to determine a pulse rate of the spontaneous pulse representation derived from the physiological signal and adjust a compression rate at which the compression robot automatically delivers CPR chest compressions to the person to a new compression rate that differs from the determined pulse rate by at least 10%. The electronic processor may be programmed to control the compression robot to stop the automatic delivery of CPR chest compressions to the person if the spontaneous pulse representation satisfies a return of spontaneous circulation (ROSC) criterion, or to send an audiovisual message to the caregiver that a spontaneous pulse representation satisfying a ROSC criterion has been detected so the caregiver can decide how to proceed next.

In accordance with another illustrative example, a control device is disclosed for controlling a compression robot configured to automatically deliver CPR chest compressions to a person. The control device comprises an electronic processor programmed to: determine a pulse rate of the person from a physiological signal measured for the person; and adjust an initial compression rate at which the compression robot automatically delivers CPR chest compressions to the person to a new compression rate that differs from the determined pulse rate by a larger amount than the initial compression rate differs from the determined pulse rate. In some embodiments, the initial compression rate differs from the determined pulse rate by less than 10% and the new compression rate differs from the determined pulse rate by at least 10%. The electronic processor may be programmed to determine the pulse rate by operations including: generating a compressions signal component by fitting to the physiological signal a harmonic series whose fundamental frequency is set to the compression rate; generating a compressions- free signal component representing the spontaneous pulse by subtracting the compressions signal component from the physiological signal; and determining the pulse rate as a fundamental frequency of the compressions-free signal component.

In accordance with another illustrative example, a control device is disclosed for controlling a compression robot configured to automatically deliver CPR chest compressions to a person. The control device comprises an electronic processor programmed to determine a spontaneous pulse representation for the person from a physiological signal measured for the person, and control the compression robot to stop the automatic delivery of CPR chest compressions to the person in response to the spontaneous pulse representation satisfying a return of spontaneous circulation (ROSC) criterion, or to send an audiovisual message to the caregiver that a spontaneous pulse representation satisfying a ROSC criterion has been detected so the caregiver can decide how to proceed next in response to the spontaneous pulse representation satisfying a ROSC criterion. The electronic processor may be programmed to determine the spontaneous pulse representation by operations including: generating a compressions signal component by fitting to the physiological signal a harmonic series whose fundamental frequency is set to the compression rate; and generating the spontaneous pulse representation by subtracting the compressions signal component from the physiological signal. The ROSC criterion may include a spontaneous pulse representation target criterion threshold (TROSCY ^e-g _> ^a minimum required pulse amplitude threshold, a minimum required pulse rate threshold, a maximum allowed variation in pulse interval threshold, or a combination thereof. The electronic processor may be programmed to control the compression robot to stop the automatic delivery of CPR chest compressions to the person one of: (1) immediately after the spontaneous pulse representation is determined to satisfy the ROSC criterion; and (2) a predefined time interval after the spontaneous pulse representation is determined to satisfy the ROSC criterion. Alternatively, the electronic processor may be programmed to send an audiovisual message to the caregiver that a spontaneous pulse representation satisfying a ROSC criterion has been detected so the caregiver can decide how to proceed next.

One advantage resides in providing more timely detection of return of spontaneous circulation (ROSC) during automated cardiopulmonary resuscitation (CPR).

Another advantage resides in providing effective detection of a spontaneous pulse during CPR chest compressions. Another advantage resides in providing effective detection of a spontaneous pulse during CPR chest compressions even when the pulse is closely synchronized with the compressions.

Further advantages of the present invention will be appreciated to those of ordinary skill in the art upon reading and understanding the following detailed description. It will be appreciated that a given embodiment may provide none, one, two, or more of these advantages.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention may take form in various components and arrangements of components, and in various steps and arrangements of steps. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention.

FIGURE 1 diagrammatically illustrates an automated cardiopulmonary resuscitation (CPR) device as disclosed herein.

FIGURE 2 diagrammatically illustrates an embodiment of the compression component separator of the automated CPR device of FIGURE 1.

FIGURE 3 diagrammatically illustrates an embodiment of the compression rate setting controller of the automated CPR device of FIGURE 1.

DETAILED DESCRIPTION

With reference to FIGURE 1, an automated cardiopulmonary resuscitation (CPR) device includes a CPR compression robot 10 comprising a motor-driven piston or other compression actuator 12 configured to be secured to the front of a torso of a person in cardiac arrest (not shown) by a harness or housing 14 that wraps around the torso. The harness or housing 14 houses various internal components not visible in FIGURE 1, such as a motor for driving the compression actuator 12 and a microprocessor or microcontroller and ancillary electronics programmed to operate the compression actuator 12 to deliver CPR chest compressions at a programmed compression rate and compression force or depth, optionally with periodic interruptions of the compressions to deliver one or two rescue breaths (i.e. ventilations). Although not shown in FIGURE 1, it is contemplated for the automated CPR device to also include an air compressor and suitable electronics synchronizing with the CPR compression robot 10 to deliver the ventilations (alternatively, an emergency responder can deliver manual breaths during the compression interruptions).

The automated CPR device further includes a physiological sensor configured to measure an arterial blood characteristic whose value exhibits variations correlating with arterial blood pressure pulsations caused by a spontaneous pulse and/or by CPR chest compressions. Two examples of such a physiological sensor are shown in FIGURE 1 to provide illustrative examples. One suitable physiological sensor is a photoplethysmography (PPG) device 20. The illustrative PPG device 20 is configured as a clip (e.g. spring-biased to close) for connection to a fingertip, earlobe, or other anatomy. The PPG device 20 is configured to measure a transmission PPG signal 22 indicating transmission through the tissue (e.g. fingertip or earlobe portion) probed by the PPG device 20. Alternatively, a PPG device configured to operate in reflection mode, or a non-contact PPG imaging device such as a camera are contemplated. The measured optical transmission varies as a function of arterial blood volume in the probed tissue due to light absorption by arterial blood, so that the PPG signal 22 exhibits temporal variations correlating with arterial blood pressure pulsations. Optionally, the PPG device 20 may be a dual-wavelength device known as a "pulse oxymeter" that is also configured to measure saturation of peripheral oxygen (Sp0₂) based on ratioing of the PPG signals at the two wavelengths.

An invasive arterial blood pressure (iABP) device 30 is also illustrated in FIGURE 1 as another example of a physiological sensor configured to measure an arterial blood characteristic whose value exhibits variations correlating with arterial blood pressure pulsations caused by a spontaneous pulse and/or by CPR chest compressions. The iABP device 30 generates an iABP signal 32 that can be quantitatively correlated with arterial blood pressure (both in terms of waveform and amplitude). The illustrative iABP sensor 30 comprises a saline (or other) fluid- filled invasive arterial line 34 terminated at one end by an arterial cannula or catheter configured for insertion into an artery of the person in cardiac arrest. The opposite end of the arterial line 34 is connected with a pressure transducer 36 (preferably with automatic flushing). A pressure bag 38 is installed to maintain a pressurized fluid column in the arterial line 32. The iAPB transducer 36 outputs the iABP signal 32 that can be quantitatively correlated with arterial blood pressure (both in terms of waveform and amplitude). An alternative for the iABP device with a fluid-filled line can be a catheter with a micro-tip (pressure) transducer to obtain a physiological signal of the person which reflects heart beats.

In general, only one physiological sensor configured to measure an arterial blood characteristic is needed. Thus, the automated CPR device of FIGURE 1 may include the PPG 20 but omit the iABP 30; or alternatively may include the iABP 30 but omit the PPG 20. In other embodiments both the PPG 20 and iABP 30 are provided. Moreover, these are merely illustrative examples, and replacing the illustrative sensors 20, 30 by a different physiological sensor configured to measure an arterial blood characteristic is further contemplated.

As previously noted, the PPG signal 22 or iABP signal 32 will exhibit temporal variations corresponding to CPR compressions performed by the compression robot 10, and will exhibit temporal variations corresponding to a spontaneous pulse generated by the heart once it has resumed beating. If both CPR compressions and a spontaneous pulse are present, then the PPG or iABP signal 22, 32 will exhibit a superposition of pulsations due to both chest compressions and spontaneous heart beats. The automated CPR device of FIGURE 1 further includes a compression component separator 40 which processes the PPG or iABP signal 22, 32 to separate out a compressions signal component 42 and a compressions-free signal component 44, the latter being a representation of the spontaneous pulse (if any).

To perform the separation, the compression component separator 40 further receives a compressions reference from which compression rate (and optionally also compression phase) information is obtained. In one embodiment, compression rate information 50 is communicated from the compression robot 10 to the compression component separator 40. This compression rate information 50 may, for example, be in the form of a digital value indicating the programmed compression rate at which the compression robot 10 is delivering compressions, or may be in the form of a control signal generated by the microprocessor or microcontroller of the compression robot 10 that cycles with the delivered CPR compressions, or so forth. In another embodiment, an automated external defibrillator (AED) 52 or more advanced monitor- defibrillator is connected to the torso of the person under cardiac arrest by defibrillator pads 54, and the AED 52 is programmed to measure trans-thoracic impedance (ΤΉ) between the defibrillator pads 54. In this embodiment, a TTI signal 56 is input from the AED 52 to the compression component separator 40. The ΤΉ signal 56 exhibits cycling corresponding to the CPR chest compressions delivered by the compression robot 10. This embodiment leverages the AED 52, which is connected to the unconscious person under some emergency response protocols in order to deliver defibrillation shock(s) if it is determined that the person is in ventricular defibrillation, in order to perform the auxiliary function of providing the TTI signal 56 from which CPR compression rate can be determined.

Either one or both of the compressions signal component 42 and the compressions-free signal component 44 are optionally displayed on a display device 60, e.g. a medical monitoring device. In the illustrative example of FIGURE 1, a trend line 62 of the compressions signal component 42 is displayed in parallel with a trend line 64 of the compressions-free signal component 44. The compressions signal trend line 62, if displayed, can inform emergency responders of the effectiveness of CPR chest compressions delivered by the compressions robot 10. The compressions-free signal trend line 64, if displayed, can inform emergency responders if a pulse is observed which may indicate the CPR has (potentially) achieved ROSC.

With continuing reference to FIGURE 1 , a compression rate setting controller 70 receives the compressions-free signal component 44, along with a compression rate 72 received from the compression component separator 40. The compression rate 72 is, or is generated from, the compression rate information 50, 56. The compression rate setting controller 70 programs the compression rate of the compression robot 10. In general, the compression component separator 40 identifies the compressions signal component 42 of the PPG or iABP signal 22, 32 as the frequency component of the PPG or iABP signal 22, 32 at the compression rate 72 (i.e. compression frequency) and its harmonics. If a spontaneous pulse is also present, it is separated into the compression-free signal component 44 if the pulse rate is sufficiently different from the compression rate. If, however, the pulse rate is too close to the compression rate, then this approach cannot distinguish the spontaneous pulse from the CPR compressions. In general, a frequency difference of about 10% or larger is sufficient for effective operation of the compression component separator 40. The compression rate setting controller 70 operates to reduce or eliminate the likelihood that the pulse (if present) has a frequency within 10% of the compression rate 72. In one approach, the compression rate setting controller 70 performs a compression rate sweep 74 over a sweep range of (for example) 90-110 compressions/minute or 80-100 compressions/minute or 80-120 compressions/minute. Such a sweep may be performed periodically to ensure that if a pulse is present the compression rate is not by chance at the pulse rate. In an alternative or complementary approach, if a pulse is detected and its pulse rate is determined to be too close to the compression rate, then a pulse rate avoidance control 76 operates to adjust the compression rate away from the pulse rate. For example, if the pulse rate is 95 beats/minute and the compression rate is 90 compressions/minute (-5% difference), then the pulse rate avoidance control 76 operates to adjust the compression rate to (for example) 110 compressions per minute (-15% difference). As a further contemplated control mode, the compression rate setting controller 70 may include a stop compressions function 80 that instructs the compression robot 10 to stop delivering CPR compressions or prompts for user input from the emergency caregiver to determine whether compressions should be continued or not. The compression rate setting controller 70 outputs a compression rate setting control signal 82 to cause the compression robot 70 to perform the desired operation. For example, the compression rate setting control signal 82 may be a numeric value or continuous control signal programming the compression rate of the robot 10, or may be a binary value turning compressions on or off. In some embodiments, the stop compressions function 80 may (in addition to or instead of actually causing the robot 10 to switch off compressions) send a ROSC indicator signal 84 to cause the display component 60 to display a ROSC indicator 86.

With reference now to FIGURE 2, an illustrative embodiment of the compression component separator 40 is described. The approach of FIGURE 2 generally follows the compression component separation approach of Wijshoff et al, "Photoplethysmography-Based Algorithm for Detection of Cardiogenic Output During Cardiopulmonary Resuscitation", IEEE Trans. On Biomedical Engineering vol. 62 no. 3 pp. 909-921 (2015). In the illustrative embodiment of FIGURE 2, the iABP compression component separator 40 receives as inputs the PPG or iABP signal 22, 32 and the trans-thoracic impedance (TTI) signal 56 from the AED 56. Digital signal processing (DSP) 90 is performed on the TTI signal 56 to extract the CPR compression rate 72, denoted herein as R[k]. The DSP 90 optionally also extracts a CPR chest compressions phase, and/or optionally also extracts an "envelope", denoted A[k], which indicates when chest compressions are interrupted (typically to perform ventilation and/or a pulse check by manual palpation). In the illustrative example, A[k] = 0 during time intervals over which chest compressions are interrupted, and A[k] = 1 otherwise. (It is noted that the illustrative compression component separator 40 operates using digital signal processing (DSP), and as is conventional in DSP literature data samples as a function of time are indexed by time index k). The envelope A[k] may be extracted, for example, by applying a peak detector or a low-pass filter and thresholding and digitizing the peak or low-pass filtered signal. The compression rate 72 can be extracted by a technique such as a Fast Fourier Transform (FFT) or other frequency-domain DSP to detect the fundamental frequency component of the ΤΉ signal 56. Alternatively, the envelope A[k] and compression rate 72 can be determined by detecting the individual compressions in the TTI signal.

With brief reference back to FIGURE 1 , in an alternative embodiment the compression rate R[k] 72 is provided directly as the information 50 provided by the compression robot 10, in which case no DSP component 90 is needed to extract that rate.

The PPG or iABP signal 22, 32 is optionally pre-processed to facilitate extraction of the component(s) corresponding to blood circulation pulsations due to chest compressions and/or spontaneous beating of the heart. In the illustrative embodiment, the PPG or iABP signal 22, 32 is high-pass filtered using a high-pass filter 92 with a cut-off frequency of 0.7 Hz or lower to generate a high pass filtered PPG or iABP signal 232. The cut-off frequency of the high-pass filter 92 is preferably chosen to pass the fundamental frequency component due to CPR chest compressions and (if present) spontaneous pulse while removing lower frequency components that are too low to be due to CPR chest compressions or a life-sustaining pulse. For some illustrative embodiments, a cut-off frequency of 0.3-0.7 Hz is contemplated.

The illustrative embodiment of the compression component separator 40 shown in FIGURE 2 employs an adaptive algorithm that estimates the CPR chest compression component 42 in the high pass filtered signal 232 by making use of the compression rate 72 determined in real-time from the compression information 50, 56 of FIGURE 1. In illustrative FIGURE 2, the estimate of the CPR chest compressions signal component 42 is by way of an operation 94 in which a harmonic series whose fundamental frequency is set to the CPR chest compression rate R[k] 72 is fit to the high pass filtered signal 232. By subtracting this estimate 42 from the high pass filtered signal 232 using a difference operation 96, the compression-free signal component 44 is obtained, which is a representation of the spontaneous pulse. In the embodiment of FIGURE 2, the high-pass filtered signal (denoted in digitized form as S[k]) is assumed to be a summation of a spontaneous pulse component sp[/e], a compression component cmp[k], and a residual component r[k]. That is, S[k] = sp[k] + cmp[k] + r[k] . The compressions signal component 42, i.e. cmp_est [k], is modelled in the operation 94 by a harmonic signal model of N_H sinusoidal harmonic components:

N_H

cmp_est [k] = A[k] Om

+ b_m[k]sin(m(p[k]) (1) m-l where N_H is the number of harmonic components, A [k] is the CPR chest compressions envelope (A[k] = 0 during intervals over which CPR chest compressions are interrupted and A[k] = 1 otherwise), and <p[/e] is an instantaneous compression phase (in radians):

where again R[k] is the CPR chest compression rate 72 (in Hertz, i.e. compressions per second), T_s is the sampling interval (in seconds), and φ₀ is an arbitrary constant phase offset (in radians). The coefficients _m[/ ] and 0_m[/ ] of the harmonic series of Expression (1) can be estimated using a least mean-square (LMS) algorithm or other optimization algorithm to fit the harmonic model of Expression (1) to the high-pass filtered signal 232. The resulting compressions signal component 42, i.e. cmp_est[k], is input to the difference operation 96 to compute the compressions-free signal component 44 according to the difference sp_est [k] = S[k]— cmp_est [k].

Fitting the harmonic series of Expression (1) effectively fits the phase of the compressions signal component cmp_est [k] by fitting both the in-phase and quadrature harmonic coefficients c½ [/ ] and b_m [k]. As a consequence, the choice of φ₀ in Expression (2) can be arbitrary as any "error" in φ₀ is removed by the fitting. This can be more explicitly seen in the alternative formulation of the harmonic series as: cmp_est[k] = A[k] ^ P_m[k]cos 12π?η ^ R[n]T_s + ^[/ ] (3)

m-1 where in this formulation the fitted parameters of the harmonic series are now the harmonic component amplitude P_m[k] and phase 0,n[/ ] in radians.

The illustrative compression component separator 40 can detect a spontaneous pulse representation sp_est[k], i.e. the compressions-free signal component 44, during CPR chest compressions. This is a representation of the spontaneous pulse in that it has the same fundamental frequency. Furthermore, if determined from the iABP, the spontaneous pulse in ^sPest[k] has an amplitude that roughly correlates with pulse strength. However, in the case of the PPG signal 22, PPG does not provide quantitative blood pressure information, in that the amplitude of the compressions-free signal component 44 of the PPG signal 22 is not necessarily related to pulse strength. By contrast, in the case of the iABP signal 32, the amplitude of the compressions-free signal component 44 of the iABP signal 32 is a blood pressure and hence is more directly related to pulse strength. On the other hand, iABP is an invasive measurement and may not be readily available in the case of a person in cardiac arrest outside of a hospital or other in-patient medical care facility; whereas it is straightforward to clip the PPG sensor 20 to a fingertip or earlobe.

With reference to FIGURE 3, an illustrative embodiment of the compression rate setting controller 70 (other than the sweep control 74) is described. In an operation 100, the fundamental frequency component of the compressions-free signal component 44 is extracted, for example by FFT or another frequency-domain DSP, or by time-domain analysis such as detecting the individual pulses or analyzing the autocorrelation. The operation 100 outputs the fundamental frequency 102 and the amplitude 104 of the fundamental frequency component of the compressions-free signal component 44. This fundamental frequency component 102, 104 is assumed to be a representation of the spontaneous pulse; accordingly, in an operation 106 if the operation 100 fails to detect a credible pulse then the process stops. This stoppage 106 may be triggered if the operation 100 fails entirely (e.g. a call to an FFT operations returns an error) or if the amplitude 104 is too small so as to indicate the compressions-free signal component 44 is close to being flat-line, or if the fundamental frequency 102 is not a credible pulse rate (e.g. too low or too high to be physiologically plausible).

If the check operation 106 verifies that the fundamental frequency 102, 104 is a plausible spontaneous pulse representation, then the fundamental frequency 102 is interpreted as a pulse rate and its amplitude 104 is interpreted as (at least roughly) corresponding to the strength of the pulse. This detected spontaneous pulse representation can then be further processed. In an operation 110, the pulse rate 102 is compared with the compression rate 72, and if these two rates are within 10% (inclusive) of each other, then the pulse rate avoidance control 76 is invoked to adjust the compression rate of CPR compressions applied by the compression robot 10 to be more than 10% different from the detected pulse rate 102. For example, if the compression robot 10 can deliver compressions in a range of 80-120 compressions per minute, then Table 1 provides a suitable algorithm for the compression rate adjustment to ensure the compression rate differs from the pulse rate (in beats per minute, i.e. bpm) by at least 20%.

With continuing reference to FIGURE 3, in an operation 112 the amplitude 104 is taken as a (rough) metric of pulse strength and is compared with a minimum threshold T_R0SC for identifying a return of spontaneous circulation. If the amplitude 104 meets this minimum threshold in the operation 112, and if the frequency 102 exceeds a minimum rate threshold, and if the variability in spontaneous pulse interval is below a maximum threshold, then the stop compressions function 80 is executed to instruct the compression robot 10 to stop delivering CPR compressions, and/or the ROSC indicator signal 84 is sent to the display component 60 to inform emergency responders that the automated CPR device of FIGURE 1 has detected potential ROSC based on which the emergency responders can decide whether the robot should stop compressions or still continue delivering compressions. This is merely an illustrative example, and the operation 112 can employ a different ROSC criterion. For example, the ROSC criterion may be a conjunctive criterion such as the amplitude 104 being greater than a minimum threshold T_R0SC and the pulse rate 102 being within a chosen range and the variation in spontaneous pulse interval being below a variability threshold. The ROSC criterion may, in general, be parameterized by the person's age, gender, size or weight, or other factors. Table 1

If the stop compressions function 80 is executed to instruct the compression robot 10 to stop delivering CPR compressions, the stoppage may be programmed to occur immediately upon the operation 112 being satisfied so as to indicate potential ROSC, or may be delayed by a delay interval in accordance with governing CPR guidelines. For example, in some embodiments compressions are continued until a complete (2 min) CPR compression phase is finished and the emergency responder can prepare for a manual pulse check.

In another variant, upon the operation 112 being satisfied so as to indicate potential ROSC, the stop compressions function 80 is executed to instruct the compression robot 10 to stop delivering CPR compressions and the physiological sensor 20, 30 continues to measure the PPG or iABP signal 22, 32 after the compressions have stopped. After the stoppage of compressions there is no any confounding compressions signal component, and so the PPG or iABP signal 22, 32 measured after the compressions have stopped provides a more accurate verification that ROSC has been achieved. On the other hand, if the measured PPG or iABP signal 22, 32 goes to a flat-line condition after the compressions have stopped then this verification fails, i.e. ROSC has not been achieved, and the compression rate setting controller 70 suitably instructs the compression robot 10 to resume delivery of CPR compressions. This monitoring is also contemplated to be extended to cause the compression robot 10 to start compressions whenever the measured PPG or iABP signal 22, 32 goes to a flat-line condition, so as to (for example) provide nearly immediate re-start of CPR in the event of a re-arrest.

In another variant, upon the operation 112 being satisfied so as to indicate potential ROSC, the stop compressions function 80 is executed to prompt to the emergency responder that a potential ROSC has been detected and that further input is required to determine whether compressions should be stopped. In this variant, there is no closed-loop control, but the actual decision about resuming or stopping chest compressions remains with the emergency responder.

The various components of the illustrative embodiment of FIGURE 1 can be variously combined or separated. For example, in one variant embodiment a trans-thoracic impedance (ΤΉ) sensor is built into the compression robot 10 along with suitable thoracic electrodes to provide the TTI signal 56, in which case the AED 52 and defibrillator pads 54 may be omitted or not used to provide the ΤΉ signal 56. The compression component separator 40 may be implemented on a microprocessor or microcontroller of the compression robot 10 or on a microprocessor or microcontroller of the display device 60, or on another electronic processor. The compression rate setting controller 70 may be implemented on a microprocessor or microcontroller of the compression robot 10 or on a microprocessor or microcontroller of the display device 60, or on another electronic processor. If these components 40, 70 are both integrated into the compression robot 10, or are implemented as one or more electronic processors other than the display device 60, then the display device 60 may optionally be omitted. It will also be appreciated that the functionality implemented on one or more electronic processors described herein may be embodied by a non-transitory storage medium storing instructions readable and executable by one or more electronic processors. The non-transitory storage medium may, for example, comprise a magnetic disk drive or other magnetic storage medium, a flash memory or other electronic storage medium, an optical disk or other optical storage medium, or so forth.

Likewise, while each of the illustrative physiological sensors 20, 30 is shown as a discrete component, such a physiological sensor configured to measure an arterial blood characteristic may alternatively be integrated with the compression robot 10. For example, a reflection-mode PPG sensor can be built into the inside of the harness or housing 14 that wraps around the torso of the person in cardiac arrest. Alternatively, a wired reflective or transmissive PPG sensor may be integrated with the compression robot, to allow for placement at a body site convenient for pulse detection. Alternatively, a non-contact imaging PPG sensor such as a camera may also be integrated with the compression robot to detect spontaneous pulses. It is also contemplated to employ another physiological sensor other than the illustrative sensors 20, 30 to assess spontaneous pulse, such as an accelerometer mounted on the torso of the person in cardiac arrest, or a radar sensor to measure spontaneous pulse through clothing.

In some variant embodiments, the physiological sensor 20, 30 is used to automatically measure spontaneous pulse (if any) only during time intervals over which the compressions robot 10 does not deliver CPR chest compressions. In such embodiments, the information 50 from the compressions robot 10 indicates time intervals during which compressions are interrupted and the signal 22, 32 from the physiological sensor 20, 30 is used to determine the spontaneous pulse only during those interruptions. In these embodiments, the sensor signal 22, 32 is the spontaneous pulse representation (i.e. there is no need to separate out any compression signal component), and the compression rate setting controller 70 suitably performs the operation 100 on the sensor signal 22, 32 directly (without the separation processing loop 94, 96 of FIGURE 2, and optionally without the high pass filter 92) to determine the pulse rate 102 and (correlative) amplitude 104, which can then be used to perform the ROSC indicator assessment operation 112 and to invoke the stop compressions function 80 as appropriate. This approach advantageously rapidly detects ROSC and promptly automatically stops CPR chest compressions, albeit with a possible delay of up to one compressions block (e.g. 30 compressions in accord with some CPR guidelines).

Where the various signals 22, 32, 50, 56, 42, 44, 72, 82, 84 are transferred between different components, the signals may be communicated via any suitable communication pathway or media, e.g. via wired links, wireless links, or so forth.

Various approaches, in addition to or alternative to those already described, may be used to invoke the pulse rate avoidance control 76. For example, in one illustrative approach if CPR is delivered in 2 min blocks of 30 compressions alternated by 2 ventilations (30:2 CPR), the pauses for ventilations can be used to perform a rapid pulse detection and pulse rate determination. If a pulse is detected and the detected pulse rate in such a pause is close to the current compression rate, then the compression rate is adjusted, for example in accord with Table 1, to facilitate monitoring of the spontaneous pulse during subsequent ongoing compressions.

In another illustrative approach, if CPR is delivered in continuous blocks of 2 min and presence of a spontaneous pulse is detected, e.g., from a change in the shape of the PPG or iABP signal 22, 32, but the pulse rate cannot be determined explicitly because it is too close to the compression rate, then the pulse rate avoidance control 76 adjusts the compression rate to facilitate monitoring of the spontaneous pulse during ongoing compressions.

In another illustrative approach, if the presence of a spontaneous pulse and its pulse rate are detected after a 2 min block of CPR, but the spontaneous pulse is considered too weak to qualify as ROSC (i.e., at the operation 112), CPR is continued. In this subsequent phase of CPR the automated chest compression rate can be slightly adjusted, e.g. using the sweep control 74, to facilitate tracking the spontaneous pulse and its pulse rate, based on the pulse rate measured during the preceding pulse check. In this way, better decisions can be made during compressions.

The invention has been described with reference to the preferred embodiments. Modifications and alterations may occur to others upon reading and understanding the preceding detailed description. It is intended that the invention be constructed as including all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims

CLAIMS:

1. A cardiopulmonary resuscitation (CPR) device comprising:

a compression robot (10) configured to automatically deliver CPR chest compressions to a person;

a physiological sensor (20, 30) configured to measure a physiological signal (22, 32) of the person; and

an electronic processor (40, 70) programmed to (i) derive a spontaneous pulse representation from the physiological signal and (ii) control the compression robot based on the spontaneous pulse representation.

2. The CPR device of claim 1 wherein the electronic processor (40, 70) is programmed to:

(i) (a) determine a pulse rate of the spontaneous pulse representation derived from the physiological signal; and

(ii) (a) adjust a compression rate at which the compression robot (10) automatically delivers CPR chest compressions to the person to a new compression rate that differs from the determined pulse rate by at least 10%.

3. The CPR device of any one of claims 1-2 wherein the electronic processor (40, 70) is programmed to (i) derive the spontaneous pulse representation from the physiological signal (22, 32) by operations including:

generating a compressions signal component (42) by fitting to the physiological signal a harmonic series whose fundamental frequency is set to a compression rate at which the compression robot (10) automatically delivers CPR chest compressions to the person; and

generating a compressions-free signal component (44) representing the spontaneous pulse by subtracting the compressions signal component from the physiological signal.

4. The CPR device of claim 3 wherein the electronic processor (40, 70) is programmed to high-pass filter the physiological signal (22, 32) before performing the generating operations.

5. The CPR device of any one of claims 1-4 wherein the physiological sensor (20, 30) configured to measure a physiological signal (22, 32) of the person comprises:

a photoplethysmography (PPG) sensor (20) configured to measure a PPG signal (22) of the person.

6. The CPR device of any one of claims 1-4 wherein the physiological sensor (20, 30) configured to measure a physiological signal (22, 32) of the person comprises:

an invasive arterial blood pressure (iABP) sensor (30) configured to measure an iABP signal (32) of the person.

7. The CPR device of any one of claims 1 -6 wherein the electronic processor (40, 70) is programmed to (ii) control the compression robot (10) to stop the automatic delivery of CPR chest compressions to the person if the spontaneous pulse representation satisfies a return of spontaneous circulation (ROSC) indication criterion.

8. The CPR device of claim 7 wherein the ROSC criterion includes at least one of a spontaneous pulse representation amplitude minimum threshold, a spontaneous pulse representation rate range or a spontaneous pulse interval variability maximum threshold (T_R0SC).

9. The CPR device of any one of claims 1-8 wherein the compression robot (10) comprises a motor-driven compression actuator (12) configured to be secured to the front of a torso of the person by a harness or housing (14) that wraps around the torso.

10. The CPR device of any one of claims 1 -9 wherein the electronic processor (40, 70) includes at least one of (A) a microcontroller or microprocessor of the compression robot (10) and (B) a microcontroller or microprocessor of a medical monitoring device (60).

11. A control device for controlling a compression robot (10) configured to automatically deliver CPR chest compressions to a person, the control device comprising: an electronic processor (40, 70) programmed to:

determine a pulse rate of the person from a physiological signal (22, 32) measured for the person; and

adjust an initial compression rate at which the compression robot automatically delivers CPR chest compressions to the person to a new compression rate that differs from the determined pulse rate by a larger amount than the initial compression rate differs from the determined pulse rate.

12. The control device of claim 11 wherein the initial compression rate differs from the determined pulse rate by less than 10% and the new compression rate differs from the determined pulse rate by at least 10%.

13. The control device of any one of claims 11-12 wherein the electronic processor (40, 70) is programmed to determine the pulse rate by operations including:

generating a compressions signal component (42) by fitting to the physiological signal (22, 32) a harmonic series whose fundamental frequency is set to the compression rate;

generating a compressions-free signal component (44) representing the spontaneous pulse by subtracting the compressions signal component from the physiological signal; and

determining the pulse rate as a fundamental frequency (102) of the compressions-free signal component.

14. The control device of any one of claims 11-13 wherein the physiological signal (22, 32) is a photoplethysmography signal (22) or an invasive arterial blood pressure (iABP) signal (32).

15. The control device of any one of claims 11-14 wherein the electronic processor (40, 70) is further programmed to one of:

control the compression robot (10) to stop the automatic delivery of CPR chest compressions to the person if a spontaneous pulse representation derived from the physiological signal (22, 32) satisfies a return of spontaneous circulation (ROSC) criterion, or prompt to the caregiver that a potential ROSC has been detected so the caregiver can determine whether to stop or continue chest compressions.

16. The control device of claim 15 wherein the ROSC criterion includes at least one of a spontaneous pulse representation amplitude minimum threshold, a spontaneous pulse representation rate range, or a spontaneous pulse interval variability maximum threshold (T_R0SC).

17. A control device for controlling a compression robot (10) configured to automatically deliver CPR chest compressions to a person, the control device comprising:

an electronic processor (40, 70) programmed to:

determine a spontaneous pulse representation for the person from a physiological signal (22, 32) measured for the person; and

control the compression robot to stop the automatic delivery of CPR chest compressions to the person in response to the spontaneous pulse representation satisfying a return of spontaneous circulation (ROSC) criterion.

18. The control device of claim 17 wherein the electronic processor (40, 70) is programmed to determine the spontaneous pulse representation by operations including:

generating a compressions signal component (42) by fitting to the physiological signal (22, 32) a harmonic series whose fundamental frequency is set to the compression rate; and

generating the spontaneous pulse representation (44) by subtracting the compressions signal component from the physiological signal.

19. The control device of any one of claims 17-18 wherein the ROSC criterion includes a spontaneous pulse representation amplitude minimum threshold, a spontaneous pulse representation rate range and / or a spontaneous pulse interval variability maximum threshold (TROSC)-

20. The control device of any one of claims 17-19 wherein the electronic processor (40, 70) is programmed to control the compression robot (10) to stop the automatic delivery of CPR chest compressions to the person one of:

- immediately after the spontaneous pulse representation is determined to satisfy the ROSC criterion; and

- a predefined time interval after the spontaneous pulse representation is determined to satisfy the ROSC criterion.