Method and system for assessing fluid responsiveness using multimodal data

[0018] reference figure 1 , Depicts a fluid reactivity assessment system 100, which illustrates a patient 10 shown lying on a bed 12, such as a typical hospital, emergency room, intensive care unit (ICU), cardiac care unit (CCU), etc. situation. According to the condition of the patient, it is also conceivable that the patient 10 may be walking, sitting in a wheelchair, sitting in a chair, etc. The patient is monitored by various medical monitoring equipment including an electrocardiogram (ECG) instrument with ECG electrodes 14 in the illustrated embodiment, a blood pressure monitor 16 (which may be, for example, a completely non-invasive Sphygmomanometer or minimally invasive arterial line), plethysmograph (PPG) 18 and capnography 20. The illustrated blood pressure monitor 16 is wrist-based; however, blood pressure monitors positioned on the upper arm or elsewhere on the patient 10 are also envisioned. If an arterial line is used to measure blood pressure, it can optionally be incorporated into an intravenous fluid delivery line or the like. The ECG 14, the arterial blood pressure (ABP) monitor 16, the PPG 18, and the capnograph 20 also include associated electronics for generating and optionally performing an ECG, blood pressure signal, organ or body (blood or air) The volume change inside and the CO in breathing gas 2 Signal processing of stress.

[0019] As in figure 1 As depicted in, the fluid reactivity assessment system 100 includes intravenous (IV) components including a fluid 26, an infusion pump 24, and a tubing 22 for delivering the fluid 26 into the patient 10. The infusion pump 24 can be implemented as a rapid infusion pump, which can be programmed via the computer system 28, can be programmed remotely, can be programmed via the user input unit 25 (touch screen/keys/touch pad, etc.). In some embodiments, the infusion pump 24 is programmed to administer a selected amount of liquid 26 to the patient 10 during a predetermined time interval, for example, a solution shock, also known as a micro solution shock. The solution shock as used in this environment allows to assess whether the patient 10 will respond to fluid resuscitation. The solution shock mentioned herein corresponds to the limited amount of liquid 26 that is infused into the patient 10 during discrete time intervals, which is different from a general liquid supply. The system 100 also includes a computer system 28 that communicates with an infusion pump 24, various electrodes 14, a monitor 16, a plethysmograph 18, and a capnograph 20 via an I/O interface 38. According to one embodiment, the I/O interface 38 can communicate with one or more of the following via a suitable communication link: various patient monitoring components 14, 16, 18, 20, a computer network (not shown), external Display device (not shown), or other suitable input electronic device or output electronic device. The system 100 may also include additional physiological monitoring equipment, such as a pulse oximeter, a thermometer, a glucose monitor, and the like.

[0020] The computer system 28 coordinates the operation of the system 100 and may be located remotely from the system 100. The computer system 28 includes at least one processor 32 and at least one program memory 30. The program memory 30 includes processor-executable instructions that, when executed by the processor 32, coordinate the operations of the computer system 28, including physiological signal analysis and processing, fluid reactivity classification, alarm generation, and the like. The processor 32 executes processor executable instructions stored in the program memory 30.

[0021] The computer system 28 also includes electronics that provide carbon dioxide monitoring 40, ABP monitoring 42, ECG monitoring 44, and PPG monitoring 46. Even though figure 1 Not depicted in, but the computer system 28 may optionally provide electronics for monitoring selected other physiological parameters (e.g., respiratory rate) based on suitable physiological input signals. The computer system 28 includes a processor 32 that can be used to control the operation of the computer system 28, help process physiological signals, execute instructions, perform calculations, help compare and store various inputs, and so on.

[0022] The computer system 28 also includes a display device 34 that communicates with the processor 32 and is configured to display one or more display information such as physiological signals, test results, patient information, vital signs, graphical user interfaces, alarms, etc.; and The user input device 36, which is, for example, a keyboard or a touch screen or a writable screen, is used to input text; and/or a cursor control device, which is, for example, a mouse, a trackball, etc., is used to convey user input information and command selections to the processor 32. In one embodiment, the display device 34 may be implemented as an integrated multi-function patient monitor. These electronic devices may be implemented in a single multi-function patient monitor (not shown). The resulting processed ECG, ABP, PPG and/or carbon dioxide signals are integrated and displayed. The display device 34 can display the measured physiological parameters, for example, ECG traces, blood pressure (BP) data, respiratory data, and so on. The display can display these parameters in various ways, for example, by displaying current values, by displaying traces of parameter values ​​according to time, and so on.

[0023] The baseline signal module 48 of processor-executable instructions stored in the memory 30 controls the reception of physiological signals from the monitors 40, 42, 44, and 46 of the patient 10 before the liquid 26 is administered during the solution shock. As discussed in more detail below, figure 2 An illustration of the baseline physiological signals 39, 41, 43, 45 received via the monitors 40, 42, 44, 46 is provided. The baseline characteristic calculator 52 of processor-executable instructions helps to derive (ie, calculate) from the baseline physiological signal 48 dynamic indications and/or characteristics from changes in ABP, ECG or PPG along the respiratory cycle, ie, pulse pressure variation Sexual (PPV), stroke volume variability (SVV) and systolic blood pressure variability (SPV). According to one embodiment, the dynamic indication can be calculated by the following formula:

[0024]

[0025] as well as

[0026]

[0027] Additional dynamic indicators and/or characteristics may include RR variability (RRV) and plethysmographic variability index (PVI). RRV may correspond to the change in the time interval between heartbeats, that is, measured by the change in the interval between hearts. RR variability (RRV) utilizes the interval between successive Rs, where R is the peak of the QRS complex, and RRV corresponds to the peak-to-peak of the successive QRS complex of the ECG wave. The plethysmographic variability index (PVI) is a measure associated with respiratory changes in the pulse oximetry waveform amplitude and is used to predict fluid responsiveness.

[0028] The baseline feature calculator 52 can also calculate various features based on the baseline physiological signal 48, such as volumetric capnography, end-tidal CO2 derived from the capnography, cardiac output derived from ABP, and so on. In the breathing cycle, the physiological signals during inhalation (inspiration) 82 and exhalation (exhalation) 80 are different. For example, in a patient who breathes spontaneously, during the exhalation (exhalation) 80 of the breathing cycle, the blood pressure increases, and during the inhalation (inspiration) 82 of the breathing cycle, the blood pressure drops. Therefore, the baseline characteristic calculator 52 is suitably configured to calculate dynamic indications and/or characteristics during the inhalation portion 82 of the breathing cycle and during the exhalation portion 80 of the breathing cycle. In one embodiment, the dynamic indications and/or features used only relate to the inhalation portion 82 of the cycle or the exhalation portion 80 of the breathing cycle.

[0029] The solution shock physiological signal module 50 of processor-executable instructions stored in the memory 30 controls the processing of the physiological signals from the monitors 40, 42, 44, 46 obtained after the administration of the liquid 26 from the IV component during the solution shock. Received to determine whether the patient 10 responds to fluid resuscitation. The infusion pump timing signal module 54 of the processor executable instructions receives the output from the infusion pump 24 during a solution shock corresponding to the automated liquid administration to the patient 10 by the pump 24. The processor-executable instruction synchronization unit 56 synchronizes the timing signal 54 of the infusion pump 24 with the solution shock physiological signal 50, so as to indicate the amount of the liquid 26 provided to the patient 10 related to the corresponding physiological signal at a given time. the amount.

[0030] The processor-executable instruction solution shock characteristic calculator 58 helps derive (ie, calculate) dynamic indications and/or characteristics from changes in ABP, ECG, or PPG along the respiratory cycle from the synchronized solution shock physiological signal 50, For example, PPV, SVV, SPV, RRV and PVI. The solution impact feature calculator 58 can further calculate various features based on the synchronized physiological signal 50 output via the synchronization unit 56, such as volume capnography, end tidal CO2 derived from capnography, cardiac output derived from ABP, and the like. The solution shock characteristic calculator 58 is suitably configured to calculate dynamic indications and/or characteristics during the inhalation portion 82 of the breathing cycle and during the expiration portion 80 of the breathing cycle that occur during the administration of the solution shock. In one embodiment, the dynamic indications and/or features used only relate to the inhalation portion 82 of the cycle or the exhalation portion 80 of the breathing cycle during the administration of the solution shock. That is, the various physiological signals 48 and 50 of the inspiratory portion 82 that reflect the associated breathing cycle of the patient 10 are used to calculate the dynamic indications and/or features 52 and 58. The breathing cycle of the patient 10 can be determined based on the operation of the capnography monitor 40.

[0031] The classification unit 60 of the processor-executable instructions uses a classification algorithm 64, for example, a probabilistic classification model. The solution impact feature and/or indication 58 and the baseline feature and/or indication 52 are input to the classification algorithm 64 to generate a liquid response The probability value is 66. According to one embodiment, the application of the classification algorithm 64 corresponds to the use of features and/or indications from the baseline 52 and the solution shock 58 as input to the logistic regression algorithm. The classification algorithm 64 may be developed or trained during previous solution shocks or preselected medical inputs based on age, weight, height, gender, ethnicity, or other medically related data. As discussed above, the features may include dynamic indications (SVV, SPV, PPV, RRV, PVI), volumetric capnography, end tidal CO2, cardiac output, and the like. As previously mentioned, dynamic indications and/or other features may correspond to physiological signals collected only during the inspiratory portion or the expiratory portion of the breathing cycle of the patient 10 being shocked by the solution. The fluid responsiveness probability value 64 represents a score indicating the likelihood that the patient 10 will actively respond to fluid resuscitation. According to one embodiment, the probability value 64 varies from 0 (ie, patient 10 does not respond) to 1 (ie, patient 10 responds to liquid).

[0032] The processor-executable instruction comparison unit 66 helps to compare the liquid reactivity probability value 64 with the threshold response value 68. The threshold response value 68 corresponds to a threshold value at or above the threshold value that the patient 10 responds to fluid resuscitation, and when it is less than the threshold value, the patient 10 does not respond to fluid resuscitation. figure 1 The fluid resuscitation evaluation system 100 provides several exemplary outputs regarding the results of the comparison unit 66. As shown in the figure, when it is determined that the liquid reactivity probability value 64 drops below the predetermined threshold 68, the processor executable instructions include an audio alarm signal 70, a visual alarm signal 72, and a control signal 74 to the infusion pump 24 (stop solution shock )Wait. The warning signal 70 can use a suitable auditory device, such as a speaker 78, and the warning signal 72 can use the display device 34 to provide a visual depiction of the patient 10’s ineffectiveness or non-responsiveness to fluid resuscitation, for example, flashing graphics, flashing Light, or various other suitable visual depiction devices or markers. The control signal 74 output to the infusion pump 24 may instruct the pump 24 to immediately stop the infusion of the liquid 26 to the patient 10. As discussed above, such liquid infusion may be harmful to the patient 10 who does not respond. Alternatively, such a signal 74 can be interpreted by the pump 24 as starting to send a suitable auditory and/or visual queue of the patient's 10 anergy to doctors, nurses, etc. When the fluid responsiveness probability value 64 of the patient 10 reaches or exceeds the threshold 66, the processor-executable instructions of the system 100 include a visual alert 76 of the patient 10's responsiveness to fluid resuscitation. As a non-limiting example, the display device 34 may generate graphics or other markings that reflect the patient's positive response to fluid resuscitation.

[0033] The computer system 28 may include a computer server, a workstation, a personal computer, a cell phone, a tablet computer, a pager, a combination thereof, or other computing device capable of executing instructions for performing the exemplary method.

[0034] According to an exemplary embodiment, the computer system 28 includes hardware, software, and/or any suitable combination thereof configured to interact with associated users, network devices, network storage devices, remote devices, etc.

[0035] The memory 30 may represent any type of non-transitory computer readable medium, such as random access memory (RAM), read only memory (ROM), magnetic or tape, optical disk, flash memory, or holographic memory. In one embodiment, the memory 30 includes a combination of random access memory and read-only memory. In some embodiments, the processor 32 and the memory 30 may be combined in a single chip. The I/O interface 38 may allow the computer system 28 to communicate with other devices via a computer network, and may include a modulator/demodulator (MODEM). The memory 30 may store data processed in the method and instructions for performing the exemplary method.

[0036] The processor 32 can be implemented in various ways, such as through a single-core processor, a dual-core processor (or more generally a multi-core processor), a digital processor, and a coordinated mathematical and/or graphics co-processor, a digital controller Wait to implement. In addition to controlling the operation of the computer system 28, the processor 32 also executes instructions and units stored in the memory 30 for executing Figure 4 Methods outlined in.

[0037] The term "software" as used herein is intended to cover any instruction set or instruction set that can be executed by a computer or other digital system in order to configure the computer or other digital system to perform the tasks that the software intends to perform. The term "software" as further used herein is intended to also cover such instructions stored in a storage medium (eg, RAM, hard disk, optical disc, etc.), and is intended to cover so-called "firmware", that is, stored in ROM Waiting for the software. Such software can be organized in various ways, and can include software components organized as libraries, Internet-based programs stored on remote servers, etc., source code, interpreted code, object code, directly executable code, etc. It is expected that the software can call system-level code or call other software residing on the server or other locations to perform certain functions.

[0038] Now go to figure 2 , An illustrative example 200 of the function of the system 100 according to an exemplary embodiment is shown. Such as figure 2 As shown, the baseline physiological signal 50 before the start of the solution shock to the patient 10 is depicted on the left. The baseline physiological signals 48 used in the illustrative example 200 include the carbon dioxide signal 39 (via the capnometer 40), the ABP signal 40 (via the ABP monitor 42), the ECG signal 43 (via the ECG monitor 44), and the PPG signal 45 (Via PPG monitor 46). These signals 48 are stored in the memory 30 and are used to generate dynamic indications and/or features 52 as discussed above. After the infusion pump 24 starts to administer the liquid systematically (ie, solution shock), the physiological signal 50 is collected for the patient 10 and the physiological signal 50 is time-synchronized with the operation of the pump 24. As shown in the figure, the expiratory phase 80 and the inspiratory phase 82 (as shaded) are distinguishable from each other and affect the signals 48, 50 collected from the patient. The expiratory phase 80 and the inspiratory phase 82 of the breathing cycle can be identified based on the carbon dioxide signal 37 or estimated based on the ABP signal 39, the ECG signal 41 or the PPG signal 43.

[0039] image 3 The collected physiological signals (48, 50) and the corresponding dynamic indications 302 derived therefrom are illustrated. As shown in the figure, SVV, SPV and PPV are derived from the ABP signal 39. As discussed above, RRV is derived from the received ECG signal 41. As discussed above, the PVI is derived from the PPG signal 43. The dynamic indicator 302 uses the respiratory changes in the ABP signal 39, the ECG signal 41, or the PPG signal 43 to evaluate fluid reactivity. In addition to the features 52 and 58, these dynamic indicators 302 are used to calculate the liquid reactivity probability value 64 via the classification unit 60.

[0040] Now go to Figure 4 , Shows a flowchart describing an exemplary implementation of a method for evaluating liquid reactivity according to an embodiment. The method starts at 400. At this time, the baseline physiological signal 48 is generated by the capnography 40, the ABP monitor 42, the ECG monitor 44, and the PPG monitor 46 from the respective electrodes 14, the monitor 16, and the plethysmography via the I/O interface 38. The device 18 and the carbon dioxide analyzer 20 receive. Then, at 402, the baseline features and indications 52 are calculated via the processor 32 or other suitable components associated with the computer system 28, for example, SVV, SPV, PPV, RRV, PVI, volumetric capnography, derived from the capnography The end tidal CO2.

[0041] At 404, a solution shock is initiated on the patient 10 via the infusion liquid 26 by the infusion pump 24 according to the preset amount, rate, and duration of administration of the liquid 26. According to one embodiment, the exemplary solution shock is a "micro solution shock" corresponding to the infusion of 100 ml of gum within 1 minute. Depending on the type of liquid being used, the appearance of the patient 10, the doctor's experience, or other factors, other micro shocks or solution shocks can be utilized according to the systems and methods described herein. An example of a frequently used solution shock can administer 500 mL of liquid within 10-30 minutes.

[0042] At 406, the physiological signals 50 of the patient 10 during the shock of the solution are received from the aforementioned input devices 14, 16, 18, 20 through the monitors 40, 42, 44 and 46. At 408, a timing signal 54 is received from the infusion pump 24 during the solution impingement. At 410, the solution shock physiological signal 50 is synchronized with the infusion pump timing signal 54 via the synchronization unit 56 or other suitable components of the computer system 28. Then, at 412, the solution impact characteristics and indications 52 are calculated via the processor 32 or other suitable components associated with the computer system 28, for example, SVV, SPV, PPV, RRV, PVI, volumetric capnography, derived from carbon dioxide The end tidal CO2. According to one embodiment, various dynamic indicators and/or features 52 and 58 may be calculated during the expiratory phase or portion 80 of the patient's breathing cycle. In such an embodiment, as in Figure 2-3 As illustrated in, the exhalation portion 80 of the breathing cycle provides greater changes to the analysis performed by the ABP monitor 42, ECG monitor 44, PPG monitor 46 and the above calculations.

[0043] Then, at 414, the classification unit 60 or other suitable component associated with the computer system 28 utilizes the classification algorithm 62, in which the benchmark feature and/or dynamic indicator 52 and the solution shock feature and/or dynamic indicator 58 are used as input . As discussed above, the classification algorithm 62 may be a logistic regression algorithm that utilizes various characteristics and/or dynamic indications 52, 58 of the patient 10 during the expiratory portion 82 or the inspiratory portion 82 of the breathing cycle. Then, at 416, a fluid reactivity probability value 64 is calculated, which represents the probability that the patient 10 will respond or not respond to fluid resuscitation. The fluid responsiveness probability value 64 represents a value between 0 and 1, where a value closer to 0 indicates that patient 10 is unlikely to respond to fluid resuscitation or will react negatively to fluid resuscitation, while a value close to 1 indicates The patient 10 responds to fluid resuscitation or will actively respond to fluid resuscitation.

[0044] Then, at 418, the fluid reactivity probability value 64 calculated for the patient 10 experiencing the solution shock is compared with a predetermined threshold response value 68. As discussed above, select a threshold response value 68 between 0 and 1, where a fluid reactivity probability value 64 greater than or equal to the threshold 68 indicates that the patient 10 will actively respond to fluid resuscitation and is less than the probability of the threshold 68 The value 64 indicates that patient 10 will react negatively or not to fluid resuscitation. The threshold 68 can be selected based on the fluid 26 being administered, the type of appearance of the injury or the condition of the patient, the amount of time allocated to the impact, the discretion of the attending physician, or a variety of other medical considerations for IV support of the patient 10. .

[0045] Therefore, at 420, it is determined whether the liquid reactivity probability value 64 is greater than the preselected threshold 68 or less than the preselected threshold 68. When an affirmative determination is made at 420, operation continues to 426, where a visual alert 76 is generated via the display device 34 or other suitable alert mechanism regarding the possibility that the patient 10 will respond positively to the fluid resuscitation. When it is determined at 420 that the liquid reactivity probability value 64 is less than the threshold 68, the operation proceeds to 422. At 422, an alarm is generated, for example, an audio alarm signal 70, a visual alarm signal 72, to warn the participant that the patient 10 may not respond to the fluid resuscitation or react negatively to the solution shock. Then, if the solution shock is still being administered, then at 424, it is terminated via the control signal 74 communicated to the infusion pump 24. According to one embodiment, the attending doctor or other caregiver responds to the alarms 70, 72 to adjust the infusion rate of the infusion pump 24 as an alternative to the shock of the termination solution.

[0046] As used herein, memory includes one or more of the following: non-transitory computer-readable media; magnetic disks or other magnetic storage media; optical disks or other optical storage media; random access memory (RAM), read-only A memory (ROM) or other electronic storage device or a collection of chips or operably interconnected chips; an Internet/Intranet server, which can retrieve stored instructions from the Internet/Intranet server via the Internet/Intranet or local area network. In addition, as used herein, a processor includes one or more of the following: one or more of a microprocessor, a microcontroller, a graphics processing unit (GPU), an application specific integrated circuit (ASIC), an FPGA, and the like. The controller includes: (1) a processor and a memory, and the processor executes computer-executable instructions on the memory to realize the functions of the controller; or (2) analog and/or digital hardware. The user input device includes one or more of the following: a mouse, a keyboard, a touch screen display, one or more buttons, one or more switches, one or more toggle switches, a voice recognition engine, etc. The database includes one or more memories. The display device includes one or more of the following: LCD display, LED display, plasma display, projection display, touch screen display, etc.

[0047] The invention has been described with reference to the preferred embodiments. Others can think of modifications and substitutions after reading and understanding the previous detailed description. This document intends to interpret the present invention as including all such modifications and substitutions as long as they fall within the scope of the claims and their equivalents.