Devices, systems and methods for encephalopathy / delirium screening and monitoring

Abstract

To provide improved systems and methods for screening and monitoring of encephalopathy / delirium.SOLUTION: Systems and methods includes the following steps of: receiving one or more signals from one or more sensing devices; processing the one or more signals to extract one or more features from the one or more signals; analyzing the one or more features to determine one or more values for each of the one or more features; comparing at least one of the one or more values or an evaluation criterion (measure) based on at least one of the one or more values with a threshold; determining a presence, absence, or likelihood of the subsequent development of delirium for a patient on the basis of the comparison; and outputting an indication of the presence, absence, or likelihood of the subsequent development of delirium for the patient.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] This application claims priority to U.S. Provisional Patent Application No. 62 / 263,325, entitled "Delirium Prediction, Screening and Monitoring," filed December 4, 2015, which is incorporated by reference in its entirety under 35 U.S.C. §119(e).

[0002] The disclosed embodiments relate to systems and methods for prediction, screening, and monitoring of encephalopathy / delirium, and more particularly, to systems and methods for determining the presence, absence, or likelihood of subsequent development of encephalopathy / delirium in a patient by signal analysis. [Background technology]

[0003] Encephalopathy (a common diagnosis known as "delirium") is a commonly underdiagnosed and highly dangerous medical condition. As we will discuss in this book, "delirium" generally refers to a syndrome that is typically diagnosed clinically based on a physician's evaluation of the patient's symptoms and diagnostic criteria. In connection with this disclosure, "encephalopathy" refers to the underlying physiological condition. As used herein, "delirium" and / or "encephalopathy" may be used in conjunction with various implementations, and the use of one is not necessarily intended to exclude the other.

[0004] Delirium (or encephalopathy) is associated with high mortality, increased risk of irreversible progressive brain decline, increased incidence of preventable complications, prolonged hospital stay, and a higher likelihood of discharge to nursing homes rather than home. They represent a severe "brain dysfunction" state and are commonly seen in a variety of hospital settings, including postoperative and elderly general medicine patients. Mortality associated with delirium is approximately 40%, as high as that of acute myocardial infarction. With costs exceeding $150 billion annually in the United States alone, delirium represents a significant burden on healthcare systems in the United States and internationally. Despite the severity of the healthcare costs and complications, there are no effective approaches to prevent and recognize delirium.

[0005] One reason delirium is underdiagnosed is the lack of a simple, objective measure to identify its impending onset. There are no devices to measure impending delirium, like an electrocardiogram for an impending heart attack or a blood test for blood glucose to monitor high risk for diabetic complications.

[0006] To date, efforts to detect delirium have relied on two main methods, both of which fall short of the practical demands of a modern hospital environment. Screening instruments, primarily based on chart review and patient interviews, have been hampered by the challenges of implementing them into clinical workflow and the failure of health care organizations to provide ongoing training in the use of such instruments. In addition, they have poor sensitivity in routine use.

[0007] Although electroencephalography (EEG) can effectively distinguish between normal brain function and delirium, it cannot logically be used to screen for delirium because it requires a skilled technician to administer (perform) the 16- to 24-lead EEG test and a sub-specialized neurologist to interpret the study (observations). This takes several hours for each patient and is nearly impossible to perform on a large number of patients in a busy hospital environment. In addition, EEG (electroencephalography) is only used to confirm its presence, not to predict the onset of delirium. Summary of the Invention [Problem to be solved by the invention]

[0008] There is a need for improved systems and methods for predicting, screening and monitoring encephalopathy / delirium. [Means for solving the problem]

[0009] Discussed herein are various devices, systems and methods for detecting, identifying or otherwise predicting encephalopathy or delirium in a patient. In various implementations, devices are utilized that detect diffuse slow wave changes, which are a hallmark of encephalopathy.

[0010] Systems and methods are described for using various tools and procedures for predicting, screening, and monitoring encephalopathy. In certain embodiments, the tools and procedures described herein can be used with one or more additional tools and / or procedures for predicting, screening, and monitoring encephalopathy. The examples described herein relate to predicting, screening, and monitoring encephalopathy for illustrative purposes only. For multi-step processes or methods, steps may be performed by one or more different parties, servers, processors, etc.

[0011] One or more computer systems may include software, firmware, hardware, or other software that is installed on the system or that causes the system to perform its operations in operation. The one or more computer programs may be configured to perform a particular task or operation by including instructions that, when executed by a data processing device, cause the device to perform an operation. One general aspect includes a system for delirium screening of a patient, including a handheld screening device including a housing, at least two sensors configured to record one or more brain signals and generate one or more values, a processor, and at least one module configured to perform a spectral density analysis on the one or more values ​​and output data indicative of an indication of the presence or absence or likelihood of subsequent onset of encephalopathy. Other embodiments of this aspect include corresponding computer systems, devices, and computer programs recorded on one or more computer storage devices, each configured to perform the operations of the above methods.

[0012] Implementations may include one or more of the following features: A system in which the module is configured to compare one or more values ​​from one or more brain signals to a threshold; A system in which the threshold is a ratio including a number of occurrences of high frequencies to a number of occurrences of low frequencies; A system in which the one or more brain signals are electroencephalogram (EEG) signals; A system in which there are two sensors; A system in which the housing includes a display; A system in which the processor is disposed within the housing; A system in which the one or more values ​​are selected from a group including high frequencies, low frequencies, and combinations thereof; A system in which the one or more values ​​are a numerical representation of a number of occurrences of each of one or more features over a period of time; A system in which the threshold is predetermined; A system in which the threshold is set based on a machine learning model; The system further includes a handheld housing including a display, at least the two sensors are in electronic communication with the housing, a processor is disposed within the housing, and the display is configured to display output data. The system further includes a validation module configured for the processor to evaluate a signal brain that converts the one or more brainwave frequencies into signal data, the validation module discarding signal data that exceeds at least one predetermined signal quality threshold. The system wherein the signal data is partitioned into windows of equal duration. The device further includes a signal processing module. The device further includes a validation module. The device further includes a threshold module. Implementations of the described techniques may include hardware, a method or process, or computer software on a computer-accessible medium.

[0013] One general aspect includes a system for assessing the presence of encephalopathy, comprising: at least two sensors configured to record one or more electroencephalogram frequencies; a processor; and at least one module configured to compare the electroencephalogram frequencies over time; perform a spectral density analysis on the electroencephalogram frequencies to establish a ratio; compare the ratio to an established threshold; and output data indicative of an indication of the absence, presence, or likelihood of subsequent onset of encephalopathy. Other embodiments of this aspect include corresponding computer systems, apparatus, and computer programs recorded on one or more computer storage devices, each configured to perform the operations of the above-described methods.

[0014] Implementations may include one or more of the following features: A system in which the threshold is pre-determined; A system in which the threshold is established based on a machine learning model; The system further includes a handheld housing with a display, the at least two sensors are in electronic communication with the housing, the processor is disposed within the housing, and the display is configured to display output data; The system further includes a validation module configured to evaluate a signal brain where the processor converts the one or more brainwave frequencies into signal data, the validation module discarding signal data that exceeds at least one pre-determined signal quality threshold; A system in which signal data is partitioned into windows of equal duration; The device further includes a signal processing module; The device further includes a validation module; The device further includes a threshold module; Implementations of the described techniques may include hardware, methods or processes, or computer software on a computer-accessible medium.

[0015] One general aspect includes a handheld device for assessing the presence or likelihood of subsequent development of encephalopathy in a patient, comprising a housing, at least one sensor configured to generate at least one electroencephalogram signal, at least one processor, at least one system memory, at least one program module configured to perform a spectral density analysis on the at least one electroencephalogram signal and generate patient output data, and a display configured to present said patient output data. Other embodiments of this aspect include corresponding computer systems, apparatus, and computer programs recorded on one or more computer storage devices, each configured to perform the operations of the above-described methods.

[0016] An implementation may include one or more of the following features: The device further includes a signal processing module. The device further includes a validation module. The device further includes a threshold module. The device further comprises a feature analysis module. The device further comprises a signal processing module. Implementations of the described techniques may include hardware, a method or process, or computer software on a computer-accessible medium.

[0017] One or more computing devices may be adapted to provide desired functionality by accessing software instructions rendered in computer-readable form. When software is used, any suitable programming, scripting, or other type of language or combination of languages ​​may be used to implement the teachings contained herein. However, software need not be used exclusively. For example, some embodiments of the methods and systems described herein may also be implemented by other circuitry, including but not limited to hard-wired logic or application-specific circuitry. Computer-implemented software combinations are similarly suitable.

[0018] While multiple embodiments are disclosed, still other embodiments of the present disclosure will become apparent to those skilled in the art from the following detailed description, which shows and describes exemplary embodiments of the disclosed apparatus, systems and methods. As will be understood, the disclosed apparatus, systems and methods can be modified in various obvious aspects without departing from the spirit and scope of the present disclosure. Accordingly, the drawings and detailed description should be regarded as illustrative in nature and not restrictive. [Brief description of the drawings]

[0019] The accompanying drawings, which are included to provide a further understanding of the invention, constitute a part of this specification, illustrate preferred embodiments of the invention, and together with the detailed description, serve to explain the principles of the invention. [Figure 1] FIG. 1 shows an overview of an exemplary screening system in use with patients. [Diagram 2] Figures 2A and 2B show an example of diffuse slowing of the brain waves as seen on an electroencephalogram. [Diagram 3] Figures 3A, 3B, and 3C show examples of diffuse slowing of the brain waves as seen on an electroencephalogram. [Figure 4A] FIG. 4A shows an exemplary system for delirium prediction, screening, and monitoring. [Figure 4B] FIG. 4B illustrates an exemplary system for a computerized embodiment of delirium prediction, screening and monitoring. [Figure 5A] FIG. 5A shows an exemplary system for prediction, screening and monitoring of encephalopathy. [Figure 5B] FIG. 5B shows an exemplary system for prediction, screening and monitoring of encephalopathy. [Figure 5C] FIG. 5C shows an exemplary system for prediction, screening and monitoring of encephalopathy. [Figure 5D] FIG. 5D shows another exemplary system for predicting, screening and monitoring encephalopathy. [Figure 6] FIG. 6 illustrates program modules and program data of a screening device according to an exemplary embodiment. [Figure 7] FIG. 7 is an overview of a method for prediction, screening and monitoring of encephalopathy, according to an exemplary embodiment. [Figure 8A] FIG. 8A is an overview of a method for prediction, screening and monitoring of encephalopathy, according to an exemplary embodiment. [Figure 8B] FIG. 8B is an overview of a method for prediction, screening and monitoring of encephalopathy, according to an exemplary embodiment. [Figure 9] Figures 9A and 9B show an example of a raw EEG signal, and Figures 9C and 9D show an example of a spectral density analysis. [Figure 10] FIG. 10 shows an example of a spectral density analysis versus time. [Figure 11] FIG. 11 illustrates an example of screening results at optimal frequency, according to an exemplary embodiment. [Figure 12] FIG. 12 shows an example of a machine learning model for identifying features of encephalopathy. [Figure 13] FIG. 13 shows an exemplary lift chart. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0020] Various embodiments disclosed or contemplated herein relate to systems, methods, and devices that can provide an objective clinical measure of encephalopathy in the clinical form of delirium. These implementations detect the presence of diffuse slowing in a patient's electroencephalogram, which is characteristic of the onset of encephalopathy. The embodiments discussed herein can detect diffuse slowing by performing a spectral density analysis on electroencephalograms recorded from a small number of distinct locations on the patient's head, thereby facilitating bedside diagnosis, such as with a handheld device. That is, various implementations can perform an algorithm that records an electroencephalogram via two or more leads placed on the patient's head, evaluates the ratio of the recorded low to high frequencies, and compares the ratio to a determined threshold to identify the onset of encephalopathy. In further embodiments, these implementations utilize additional data, such as machine learning and data from medical records, to improve diagnostic accuracy.

[0021] The disclosed systems, devices and methods relate to non-invasive point-of-care diagnosis using fewer leads than the 16, 20, 24 EEG found in the prior art. For example, as generally shown in FIG. 1, in one particular implementation, a two-lead "bispectral electroencephalography" (BSEEG) screening system 1 is used, which can be performed with a handheld screening device 10 by attaching two leads 12A, 12B to the forehead of a patient 30 for less than 10 minutes. In these implementations, two leads or channels in a BSEEG are used for illustration, but it is understood that a greater number of leads or channels are contemplated herein. In various implementations, the device 10 can display graphical and / or numerical representations 7 of useful information for use in diagnosing encephalopathy / delirium, such as last measurements 4, trends 6, signal quality 8, etc. It is understood that these representations 7 can be the result of understood graphical user interface techniques on the display 16.

[0022] EEG can have various frequencies and / or bands of frequencies. "Diffuse slowing" is a characteristic indicator of encephalopathy. Also, Figures 2A-3D show some EEG measurements from a patient experiencing symptoms of encephalopathy 2 compared to a normal control 3. As will be evident to one skilled in the art, in various conditions, EEG in encephalopathy can be characterized by "diffuse slowing" 2, meaning that slow waveforms can be observed in each channel observed. As is evident from Figures 2A-2B, this slowing is diffuse rather than localized, so that slowing (shown in Figure 2A) is observed in most, typically all, of the various electrodes from the EEG (electroencephalograph).

[0023] As shown in Figure 3A, the appearance of a more gradual wave 2 compared to the number of higher frequency waves may be a sign that the patient is developing or is more likely to develop encephalopathy.

[0024] As shown in Figures 3B-3D, diffuse slow waves can be routinely observed in all or nearly all of the EEG electrodes, so it is possible to utilize the standard number 16-24 to identify diffuse slow waves and predict the onset of encephalopathy. In these implementations, the two leads utilized in the BSEEG implementations discussed herein are sufficient to detect diffuse slow waves and, including appropriate signal processing and user interface, do not require special expertise for placement or interpretation and may be performed with the aid of a simple handheld screening device.

[0025] In FIG. 4A, one implementation of a screening device 10 is shown. The systems and methods for predicting, screening and monitoring encephalopathy disclosed herein can utilize such handheld or other portable screening devices 10. In these implementations, the screening device 10 is configured to receive signals from, for example, one or more sensors 12A, 12B. Because diffuse slow-waves are easily identifiable throughout the patient's brain, these devices 10 can use much fewer than 20 sensors, such as two, three, four, five or more sensors 12. In certain implementations, six to twenty or more sensors are used. Thus, it is understood that since fewer than twenty sensors 12 are used, the disclosed devices and systems can be easily carried and used on the patient 30, since there is no need to apply a typical prior art EEG cap (cap-type EEG electrode).

[0026] In an implementation of FIG. 4A, the one or more sensors 12A, 12B are brain sensors, such as, but not limited to, electrodes placed on the patient. In a particular embodiment, the signal may be an electroencephalograph (EEG) signal from one or more electrodes measuring the patient's brain activity. The signal may be processed to extract one or more features of the signal. The one or more features may be analyzed to determine one or more values ​​for each of the one or more features. If the patient does not currently exhibit clinical signs or symptoms of disease, the values ​​or a metric based on the one or more values ​​may be compared to a threshold to determine the presence, absence, or likelihood of subsequent development of encephalopathy.

[0027] Continuing with FIGS. 4A-4B, the screening device 10 can include a housing 14. In various implementations, the housing 14 can be handheld or otherwise portable in size. The housing 14 can have a display 16 and an interface 18, such as buttons or a touch screen, as preferred by those skilled in the art. In various implementations, the sensors 12A, 12B are connected to the device 10 via ports 22A, 22B and wires 24A, 24B, while an alternative implementation utilizes a wireless interface, such as Bluetooth. As shown in FIG. 4A, a ground lead 13 can also be provided, and the ground lead 13 can be bundled and streamlined into one of the wires 24A, 24B. In various implementations, a transmission cord 26 or other connection can be used to electrically communicate the device 10 with a server or other computing device, as discussed in connection with FIGS. 5A, 5B.

[0028] As also shown in FIG. 4B, in various embodiments, the display 16 can depict graphical and / or numerical representations 7 of useful information for use in diagnosing encephalopathy / delirium, including one or more EEGs 2A, 2B, last measurement 4, trends 6, signal quality 8, etc. In an exemplary implementation, a graphical representation of a threshold step described below, comparing the spectral density to an established threshold, is shown as the last measurement 4. It is understood that these representations 7 can include any program data 67 and can be shown to the provider as a result of understood graphical user interface techniques on the display 16. In these implementations, the EEGs 2A, 2B are derived from a sensor 12 attached to the patient and can be used to diagnose and / or predict the onset of encephalopathy in clinical or other situations. In various implementations, the housing 14 also includes a power source and computer-related components such as a microprocessor, memory, etc. In additional implementations, calculations and other processing are performed within the device housing 14, such as via program modules (described in connection with FIGS. 6, 7), although additional processing can occur elsewhere as described herein. In each of these implementations, the screening device 10 and / or system 1 includes a processor configured to determine the patient's diffuse slowing by spectral density analysis performed on fewer than 20 channels to identify diffuse slowing.

[0029] FIG. 5A illustrates an exemplary system 1 for prediction, screening, and monitoring of encephalopathy according to one implementation. In this implementation, the system 1 can include one or more screening devices 10 (e.g., screening device 1, screening device 2, ..., screening device n), which is the device 10 of FIG. 1 in a particular implementation. The screening device 10 in this implementation is operatively coupled to one or more sensors 12A-12-n. The one or more sensors 12 may be individual sensors, an array of sensors, a medical device, or other computing device, a remote access device, etc. In a particular embodiment, the one or more sensors 12 may be one or more electrodes and / or other brain function monitoring devices. The one or more sensors 12 may be directly or indirectly coupled to the patient 30 to monitor the biological signals of the patient 30.

[0030] As also shown in Figure 5A, various processing and analysis of signals received from sensors 12A, 12B, 12C, 12-n may be performed by screening device 10, as described in connection with Figures 6 and 8. In certain alternative implementations, with reference to Figures 5B-5D, screening device 10 may be used in conjunction with other computing devices.

[0031] As shown in the implementation of FIG. 5B, the screening device may be connected or otherwise interfaced with the bedside device 11. In various implementations, the bedside device 11 may be a patient monitor or other unified bedside monitoring device 11. The bedside device 11 may be used by a medical service institution to view and / or perform certain analysis or observation steps.

[0032] 5C, 5D, one or more screening devices 10 may be operatively connected directly and / or indirectly to one or more server / computing devices 42, system databases 36 (e.g., Database 1, Database 2, ..., Database n), via a network or the like. A variety of devices may be connected to the system, including, but not limited to, medical equipment, medical monitoring systems, client computing devices, consumer computing devices, provider computing devices, remote access devices, etc. The system 1 may receive one or more inputs 38 and / or one or more outputs 40 from various sensors, medical equipment, computing devices, servers, databases, etc.

[0033] 5D in some embodiments, one or more screening devices 10 may be directly coupled and / or integrated with one or more server / computing devices 42 via connections 32 and / or may be coupled to one or more server / computing devices 42 through one or more network connections 32. The screening device 10 and / or one or more servers / computing devices 42 may also be operatively connected directly and / or indirectly to one or more third party servers / databases 34 (e.g., database 1, database 2, ... database n) via a network or the like. The one or more servers / computing devices 42 may represent, for example, any one or more of a server, a personal computer (PC), a laptop, a smartphone, a tablet, or the like.

[0034] In various implementations, the connection 32 may represent any combination of, for example, hardwired connections, wireless connections, and local area networks, such as the Internet, an intranet, a wide area network, a cellular network, a Wi-Fi network, etc. The one or more sensors 12, which may include at least one processor and / or memory, may represent any set of sensors, medical equipment, or computing devices, executing applications that send data inputs to and / or receive data outputs from one or more screening devices 10 or server / computing devices 42, respectively. Such server / computing devices 42 may include, for example, one or more of desktop computers, laptops, mobile computing devices (e.g., tablets, smartphones, wearable devices), server computers, and / or the like. In certain embodiments, the input data may include analog and / or digital signals, for example, from an EEG system, other electroencephalography measurements, etc., for processing at one or more server / computing devices 10.

[0035] In various implementations, the data output 40 may include, for example, medical indications, recommendations, notifications, alerts, data, images, and / or others. Embodiments of the disclosed embodiment may also be used for collaborative projects where multiple users log in and perform various operations on the data project from various locations. Certain embodiments may be computer-based, web-based, smartphone-based, tablet-based, and / or human wearable device-based.

[0036] In other example implementations, the screening device 10 (and / or the server / computing device 42) may include at least one processor 44 coupled to a system memory 46, as shown in FIG.

[0037] Spectral density analysis from the leads can be performed by a variety of electronic and computer mechanisms. In the implementation of FIG. 6, program modules 48 are provided. These program modules can include a signal processing module 52, a feature analysis module 54, a validation module 55, a diffuse slow wave or encephalopathy determination or threshold module 56, an output module 59, and other program modules 58, such as an operating system, device drivers, etc. In various implementations, each program module 52-58 can include a respective set of computer program instructions executable by the processor 44.

[0038] 6 illustrates one example of a set of program modules, with other numbers and arrangements of program modules being contemplated as a function of any particular design and / or architecture of the server / computing device 10 and / or system 1. In an exemplary implementation, program data 50 includes signal data 60, feature data 62, validation data 63, diffuse slow wave or encephalopathy determination module 64, output data 67, and other program data 66 (e.g., data inputs, third party data, and / or otherwise configured to perform the various steps described in FIGS. 7 and 8 ).

[0039] In various implementations, the program data 50 can correspond to the various program modules 48 described above. In these various implementations, the program modules 48 and / or program data 50 may be used to record, analyze, and otherwise output output data 67 to either a display (as shown at 16 in FIG. 1A and FIG. 8B) and / or other computing devices described herein, such as the bedside monitoring device 11 in FIG. 5B, the various system databases 36 and / or third party servers or databases 34 in FIG. 5C, the system server / computing device 42 in FIG. 5D, or, in implementations where the modules are not contained within the housing 14, the screening device 10 itself and the gateway 210 in FIG. 12 or any other computing, storage, or communication device in electrical or physical communication with the output module. It will be understood that the output data 67 provides a point-of-care provider with characteristics of the patient's brainwaves to diagnose or identify symptoms of delirium or encephalopathy. For example, in certain implementations, the last measured reading of the spectral density may provide a qualitative measure of signal quality and trend, as described elsewhere herein.

[0040] 7 is an overview of an exemplary method 5 according to one embodiment for predicting, screening and monitoring encephalopathy. To achieve prediction, screening and monitoring of encephalopathy, the system and method can perform several optional steps.

[0041] One optional step is a recording step 70, in which input data, such as one or more signals, may be received and / or recorded by one or all of the sensors 12, device 10, processor 40 and / or system memory 46 shown, for example, in Figures 4A, 4B, 5A-5D and 6.

[0042] Continuing with FIG. 7, in an optional processing step 71, one or more signals can be processed, such as by program data 50 and / or signal processing module 52 shown in FIG. 6. In one particular implementation, these signals can be processed to partition the signals into windows, as also shown in FIG. 8A. The signals can also be processed to extract one or more values ​​from one or more signals and / or windows, as further described with respect to FIG. 8A. These values ​​can include features of the raw signals from the EEG, in certain embodiments. In various implementations, the system 1 can orient these values ​​in a way that improves the accuracy of the system, as further described with respect to FIGS. 11A and 11B.

[0043] Continuing with Figure 7, an optional analysis step 72 may be performed in which one or more values ​​may be analyzed to determine particular features of the signal or window. In one particular implementation, this analysis step may include performing a Fast Fourier Transform 100 to generate or otherwise compare feature data (shown at 64 in Figure 6).

[0044] In another optional step, a validation step 73 can be performed. In these implementations, as shown in FIG. 8A, the spectral density 102 and / or other raw signal values ​​104A, 104B or other features can be used to compare individual readings from a partitioned signal window (generally shown as 84A, 84B) for exclusion or inclusion from use in subsequent steps. In various implementations, this step can be performed by the validation module 55 and validation data 63. In various embodiments, the validation step 73 can be performed in parallel with other steps described herein or otherwise in repeated iterations, utilizing an error collection algorithm, until the resulting signal data is ready for subsequent analysis. For example, as further described in FIG. 8A, the validation step 73 can be used to ensure that all readings above or below a certain measured predetermined error value are excluded from further processing.

[0045] Continuing with FIG. 7, a diffuse slowing or encephalopathy determination step, or threshold step 74, may be performed. In this step, an assessment criterion (scale) based on at least one of the one or more values ​​or features, or at least one of the one or more values, may be compared to an established diffuse slowing threshold (shown at 65 in FIG. 8), actionable via diffuse slowing threshold data (shown at 64 in FIG. 6). In certain embodiments, an indication of the presence, absence, or likelihood of subsequent development of encephalopathy in the patient based on the threshold comparison may be determined. For example, as discussed in the illustrative example below, using the average of the 3Hz / 10Hz readings as threshold data 64 and a threshold 65 of 1.44 (positive≧1.44 and negative<1.44), the screening device 10 may graphically indicate positive and / or negative readings with or without displaying the threshold 65. It is understood that the threshold 65 may be modified over time as a result of improvements and enhancements to the system, as well as based on additional data 66 and other factors.

[0046] In an exemplary implementation, a graphical representation of the threshold steps is displayed on the screening device 10 or other monitoring system, and a comparison of the spectral density to the established threshold is shown as the final measurement 4. Each of these optional steps is discussed in further detail below in connection with the embodiments disclosed herein.

[0047] As shown in Figure 8A, in one exemplary implementation of the system 1, a patient 30 is monitored by sensors 12A, 12B, such as with the screening device 10 of Figure 1. In this implementation, the sensors 12 generate one or more signals 80A, 80B from the patient 30. The one or more signals 80A, 80B may be analog and / or digital signals. The one or more sensors 12A, 12B may be separate from the system that receives the one or more signals, or may be integral to the system that receives the one or more signals. While the currently described system 1 is related to collecting EEG signals, in alternative implementations, one or more sensors 12A, 12B may measure, detect, determine and / or monitor one or more of the following physiological conditions: heart rate, pulse rate, EKG (electrocardiogram), heart rate variability, respiration rate, skin temperature, exercise parameters, blood pressure, oxygen levels, cardiac temperature, extracorporeal heat flow, galvanic skin response (GSR), electromyogram (EMG), electrooculogram (EOG), body fat, hydration levels, activity levels, oxygen consumption, glucose or blood sugar levels, body position, pressure on muscles and bones, ultraviolet (UV) absorption, and the like.

[0048] In various implementations, as described with respect to Figures 1A-6, system 1 includes one or more signal processing devices that may be located within screening device 10 or elsewhere. Processing may determine one or more features by analyzing one or more signals to look for signal information corresponding to particular signal features. In certain embodiments, one or more signals may be processed to determine the presence of one or more high frequencies and / or one or more low frequencies. The one or more features may be, for example, but not limited to, a type of wave, such as high frequency and / or low frequency.

[0049] In the FIG. 8A implementation, signals 80A, 80B are received through a first 82A and a second 82B channel for processing, as described above. Other implementations are possible. The methods can also be integrated into an 8-bit or 16-bit embedded device environment, or a more robust environment. The methods can be integrated into an existing hospital patient workflow.

[0050] As also shown in FIG. 8A, in certain implementations, signals 80A, 80B may be processed into windows of equal sequential divisions of the signal (generally shown as 84A and 84B). In the illustrated implementation, a signal duration of 10 minutes is shown, and the signal is windowed into 10 separate 1-minute windows (generally shown as 84A and 84B), although it is understood that in other implementations, other signal durations and numbers of windows may be used. In various implementations, the signal may be less than 1 second, may be 1 hour or more, or may be any number of minutes. Similarly, there may be any number of windows, and the windows may range in time intervals (divisions) from 1 second to a few minutes or more.

[0051] In the implementation of Figure 8A, after the signals 80A, 80B are segmented, several optional processing and analysis steps can be performed as described with respect to Figure 7. In various implementations, the values ​​86A, 86B are extracted and various optional steps can be performed on the values ​​86A, 86B to identify raw data features 104A, 104B, such as some or all of the minimum / maximum amplitudes 88A, 88B, mean amplitudes 90A, 90B, interquartile range (IQR) amplitudes 92A, 92B, mean deviation of amplitudes 94A, 94B and / or signal entropy 96A, 96B that can be established for the respective signals 80A, 80B. Other raw data features 104A, 104B can be identified, as would be apparent to one of ordinary skill in the art.

[0052] In the analysis step, a spectral density analysis 102 may be utilized to identify diffuse slowing. As also shown in FIG. 8A, in certain embodiments, a fast Fourier transform 100 may be performed on one or more signals 80A, 80B to identify additional features 104A, 104B. The fast Fourier transform 100 may be used alone or in combination with other analysis tools to evaluate the spectral density. A spectral density analysis 102 may be performed on one or more signals to determine one or more spectral density features 104A, 104B, such as low frequency density 105A, high frequency density 105B, and a ratio (ratio) of low frequency density to high frequency density 105C. These features 104A, 104B from the spectral density analysis 102 may be used alone or in combination with various values ​​104A, 104B derived from the raw signals by using various optional steps and combinations to predict and / or diagnose encephalopathy.

[0053] In a particular implementation, a spectral density analysis 102A, 102B may be performed on each of one or more signals 80A, 80B, such as the illustrated EEG signal, to distinguish between different patient conditions. In a particular embodiment, the spectral density analysis 102A, 102B may provide values ​​104A, 104B including a ratio 105C of high frequency brain electrical activity to low frequency brain electrical activity. For example, to establish diffuse slowing, a ratio of approximately 10 Hz signals to approximately 2 Hz, 3 Hz, or 4 Hz signals may be compared. One or more bands or windows within one or more signals 80A, 80B may be identified for use in the systems and methods described herein.

[0054] In certain implementations, a validation step 73 can be performed (shown in box 106). In these implementations, the spectral density 102 and / or other raw signal values ​​104A, 104B can be used to compare individual readings from a partitioned signal window (generally shown as 84A, 84B) for exclusion or inclusion from the analysis. For example, in certain implementations, a correction algorithm 108 can be performed. In one implementation, the error collection algorithm 108 performs a number of optional steps. In one optional step, various values ​​104A, 104B above or below a certain predetermined error threshold in a given window are discarded (110). In another optional step, if the IQR and / or density ratio 105C are within a predetermined proximity 112, these window signals are retained for aggregation and recombination 116, as described below. Other optional steps are possible.

[0055] In these implementations, an optional additional recombination step 116 may be performed in which windows 84A, 84B that were not removed as a result of the validation step 73 (box 106) may be combined such that the values ​​104A, 104B, 104A, 104B from those windows are aggregated as the diffuse slowing threshold data 64.

[0056] Using the diffuse slowing threshold data (shown as 64 in FIG. 6), a diffuse slowing or encephalopathy thresholding step 74 is performed as further illustrated in FIGS. 9A-13. In these implementations, the aggregated threshold data 64 can be compared to a threshold 65 established by the threshold module 56. In various implementations, data from other sources 66, such as electronic medical records, is also compared to the threshold 65 to establish an output 67 that can be presented graphically, for example, in connection with FIGS. 5A-5D, or in any of the other devices and systems described herein. One such example is a display of the output data 67, which can include last measurement 4, trend 6, quality 8, power 75, and other graphical representations, as seen in FIGS. 9A-9D.

[0057] As shown in Figures 2A-3D, the EEG in encephalopathy is characterized by "diffuse slow wave", which means that the electrodes, preferably all electrodes, from the EEG show slow waveforms. As mentioned above, the EEG may have various frequencies and / or bands of frequencies. Diffuse slow wave may mean that slower waves, lower frequencies, are seen across most, if not all, electrodes of the EEG. The appearance of slower waves compared to the number of higher frequency waves may be a sign that the patient has or is more likely to develop encephalopathy. To detect diffuse slow wave, two leads (and ground) are appropriate and may not require special expertise for placement or interpretation, with appropriate signal processing and user interface.

[0058] In one particular embodiment, referring to Figures 9A, 9B, a spectral density analysis 102A, 102B is performed on one or more values ​​104A, 104B, which may be the number of high frequencies in one or more signals and / or the number of low frequencies in one or more signals. As shown in Figure 9A, each channel 82A, 82B may be analyzed by the steps described in connection with Figure 8A to determine "high" and "low" frequencies and other characteristics as output data 67. In various implementations, these representations may be used as any of the various forms of data discussed above in this process and may be displayed on the device display 16 (shown in Figure 4B).

[0059] In certain embodiments, the values ​​104A, 104B may be calculated over the entirety of one or more signals 80A, 80B as part of any of the steps described above. In certain embodiments, the values ​​104A, 104B may be calculated over a subset of one or more signals or a subset of time of one or more signals. For example, if one or more signals are 5 minutes in duration, the values ​​104A, 104B may be calculated over less than 5 minutes, such as 4 minutes, 3 minutes, 2 minutes, 1 minute, 30 seconds, etc. Thus, the one or more values ​​104A, 104B may be features 104A, 104B and / or values ​​104A, 104B over a given amount of time. In certain embodiments, the one or more values ​​104A, 104B may be a number of high frequencies over a period of time and / or a number of low frequencies over a period of time. In certain embodiments, the one or more values ​​104A, 104B may be a ratio of a number of high frequencies to a number of low frequencies. In certain embodiments, one or more of the values ​​104A, 104B may be a ratio of a number of high frequencies over a period of time to a number of low frequencies over a period of time.

[0060] As shown in FIGS. 8A, 8B and 9A to 9D, a device 10, a system 1 and In various implementations of method 5, the ratio of high-frequency to low-frequency waves for each channel may be plotted and compared in a diffuse slowing or encephalopathy determination step (shown at 74 in FIG. 7). In these implementations, a metric based on at least one of one or more values ​​or features or at least one of one or more values ​​included in threshold data 64 may be compared to a diffuse slowing threshold 65 established via module 56. In certain embodiments, based on the comparison, the presence or likelihood of subsequent development of encephalopathy for the patient may be determined. In various implementations, frequencies are compared over time for each signal and the results are analyzed. As described further below, FIGS. 9A and 9B show the raw EEG channel signals over time for each of the respective channels. FIGS. 9C and 9D show the spectral density analysis results 102 for each channel.

[0061] Experimental results are demonstrated and conclusions are given in the accompanying examples.

[0062] Example 1: Screening for Encephalopathy with BSEEG Compared with Clinical Diagnosis of Delirium

[0063] In this example, a pilot study was conducted utilizing over 80 patients aged 65 years and older, with and without clinical delirium, to compare EEG signals obtained by screening device 10, system 1 and method 5. Baseline cognitive function was assessed using the Montreal Cognitive Assessment (MoCA).

[0064] In this example, patients were screened for the presence of delirium using the Method for Assessment of Delirium in the ICU (CAM-ICU). Following the assessment, an EEG reading was taken using the currently described BSEEG device, system and method, i.e., two EEG leads were placed on the patient's forehead to obtain two channels of signal, one for each hemisphere, from the left and right over the course of 10 minutes. A ground lead was also used. This process was repeated twice daily during the patients' hospital stay for up to seven days, after which point the study was terminated if no change in mental status was observed. If a change in mental status was observed, the change was monitored for an additional period of time.

[0065] In this example, the quality of the EEG signal from the presently described screening device was compared simultaneously with EEG signals obtained from a traditional 20-lead EEG machine on the same patients, and it was established that there were no significant differences between the results.

[0066] In this example, a preliminary analysis of data from a limited number of cases with and without delirium was performed. Initial analysis showed that the currently described device, system and method clearly distinguished between these patients and detected delirium and absence of delirium in the same patient at different times. Based on the BSEEG results and the signal processing algorithm, the number of subjects correctly classified as positive (true positive) and negative (true negative) is counted. In this example, the number of cases incorrectly classified as positive (false positive) and negative (false negative) is tabulated. It is possible to calculate sensitivity and specificity at several thresholds for positive and negative results in comparison to the CAM-ICU findings.

[0067] A receiver operating characteristic (ROC) analysis was then performed to develop an algorithm for evaluating the screening device output data. The ROC process was repeated with multiple algorithms to develop an optimal algorithm with a target predictive accuracy (AUC) of 0.7 or greater (1.0 is ideal). In various implementations, this algorithm can be incorporated into a system, such as the handheld screening device 10 (e.g., as shown in FIG. 4A) or elsewhere in the system 5 for use in bedside diagnostics and system improvement, as described above.

[0068] Example 2: Evaluation of screening devices

[0069] In this example, the initial training set data set included 186 whole-patient EEG samples that were correlated with clinical or CAM (Delirium Assessment Method) evidence of delirium. These samples represented 5 positive, 179 negative, and 2 negative cases with insufficient data quality for the analyses performed and were therefore excluded from further consideration.

[0070] In this example, a 15 Hz low-pass filter was originally used, but preliminary results showed that the FFT frequency information was unevenly attenuated between positive and negative cases, so the low-pass filter was removed.

[0071] During processing of processed samples, it was observed that a 4 second window was sufficient to demonstrate good results, and in this example, windows containing high amplitude peaks were excluded using a threshold of 500 μV, as shown, for example, in Figures 9A and 9B.

[0072] 9C and 9D also show the spectral density of channels whose intensity (in W / Hz) can be compared with each frequency (in Hz) to establish the low / high frequency ratio. It was observed that using a 3Hz / 10Hz ratio gave favorable prediction results in this training set, as opposed to a 2Hz / 10Hz or 4Hz / 10Hz ratio.

[0073] In FIG. 10, system 1 and method 5 can be used to establish a spectral density analysis of time series 130 to compare delirious 132 to "normal" 134 wave frequencies over time. As evident from delirious subject 139, the system can identify an increase in the ratio of low (3 Hz) to high (10 Hz) waveforms compared to the control (subject 122), thereby indicating the presence of encephalopathy. In various implementations, these results can be used to prospectively identify the onset of encephalopathy by establishing spectral density compared to a population threshold.

[0074] As shown in Figure 11, CAM-ICU (or other clinically identified patients) In contrast, the threshold 65 was set at 1.44 (positive ≥ 1.44 and negative < 1.44). Table 1 shows the performance indicators that use the average of 3Hz / 10Hz as the threshold data64. As shown in.

[0075] [Table 1]

[0076] It will be appreciated that numerous additional examples may be provided.

[0077] Example 3: Machine Learning

[0078] As shown in FIG. 12, in a particular implementation, a machine learning model (box 200 (machine learning model) A machine learning module (MLM) may be used to ascertain features of delirium / encephalopathy and refine thresholds (65 in FIGS. 6-8) to improve the accuracy of diagnosis, as well as other systems described herein. In these implementations, models are used to correlate personal and population patient data (generally shown in box 202) in a computational (computer) (box 204). In general, various machine learning approaches may be coded for execution on the screening device 10, server / computing device 42, database 36, third party server 34, or other computing or electronic storage device in operative communication with the screening device 10 and / or sensor 12 (also shown in the implementations of FIGS. 5A-5D).

[0079] The model can be run on recorded or otherwise observed data (box 202) from patient 30 (e.g., the spectral density analysis 102, output data 67, and other values ​​described in connection with FIG. 8A), as well as other recorded data, such as data from an electronic medical record (EMR - box 201).

[0080] In various implementations, the EMR data 201 may include, but is not limited to, one or more of the following physiological conditions: heart rate, pulse rate, EKG, heart rate variability, respiration rate, skin temperature, motion parameters, blood pressure, oxygen level, cardiac temperature, extracorporeal heat flow, GSR, EMG, EEG, EOG, body fat, water level, activity level, oxygen consumption, glucose or blood sugar level, body position, pressure on muscles and bones, ultraviolet (UV) absorption, etc. Observed data output from these combined EMR and other systems described herein may also be provided to the EMR, a separate patient monitoring system, a graphical user interface on the device, etc.

[0081] Thus, various systems and methods using the machine learning model (box 200) can transmit and / or receive information from various computing devices as well as patient EMRs via gateway 210 or other connectivity mechanisms for delirium monitoring, screening, or prediction. In certain implementations, the systems and methods can utilize EMR data to improve the accuracy of delirium monitoring, screening, or prediction performed in connection with screening device 10 and associated systems 1 and methods 5.

[0082] In various implementations, the patient data 202 can be loaded into any computer storage device of a computer to generate an appropriate tree algorithm or logistic regression equation. Once generated, the tree algorithm, which can take the form of a large set of if-then conditions, can be coded using any common computational language for performing tests. For example, the if-then conditions can be captured and compiled to generate a machine executable module (box 206) that upon execution receives new patient data 202 and outputs results 208, which can include calculated predictions or other graphical representations (box 208). The output can be in the form of a graph showing predicted or probability values ​​along with associated statistical indices such as p-values, chi scores, etc. In various implementations, these results 208 can be reintroduced into the learning module 200 or elsewhere to continually improve the functionality of the system, including updating various thresholds used throughout. It is understood that these implementations can also trend the respective data values ​​and readings to improve the performance of the devices, systems, and methods. In these implementations, for example, a continuous stream of trend data can be identified that can be used to provide additional optional evaluation steps, as well as trends over time. In various implementations, the model may provide additional program data (shown at 48 in FIG. 6) to improve accuracy, and may also be included in the collection as shown at 116 in FIG.

[0083] Generate better algorithms and weight variables in machine learning models (box 200) Enhanced classification and regression tree techniques can be used to further determine the significance of a given set of events, for example using classification and regression trees, random forests, boosted trees, support vector machines, neural networks, as well as other machine learning techniques previously mentioned. A lift chart showing the lift values ​​for each of these approaches is shown in Figure 13. As shown in.

[0084] Tree boosting techniques combine a set of classifier variables to achieve a final classifier. In various implementations, this is done by constructing an initial decision tree based on model development data to classify the dependent variable. For all cases in the development dataset with incorrectly classified outcomes, the weights of these cases are increased (boosted) and a new decision tree is generated to optimize the classification of outcomes based on the weights of the new cases. The incorrectly classified cases are reweighted and a new decision tree is generated. This approach is repeated, typically hundreds or thousands of iterations, until an optimal boosted tree is identified. This boosted decision tree is then applied to a validation dataset, and the cases in the validation dataset are classified as responders or non-responders. Many other known approaches may be used.

[0085] It should be noted that both the boosted tree machine learning techniques and the more sophisticated tree generation techniques can generate highly complex algorithms (containing many if-then conditions) as described above. Instead, the choice of variables and / or the relative importance of variables used as inputs to either the regression and classification tree techniques to generate the algorithm uniquely identifies the algorithm.

[0086] In this example, a machine learning algorithm (similar to the algorithm in Box 200) was developed on a cohort of 13,819 patients. In this example, variables such as laboratory values, medications, and age were utilized. The algorithm was trained on 12,461 of these patients and validated on 1,358 test cases. The delirium outcome in each patient was compared to the DOSS scale, a manually administered screening tool in which a score greater than 3 indicates delirium.

[0087] In this example, the performance of the model was as follows: true negative: 951; false positive: 62; true positive: 132. Thus, the observed error rate was 20% and the accuracy was 80%. Importantly, this model did not include the output from the screening device 10. It will be appreciated that when used in combination with the screening device 10, the accuracy can be increased beyond 80%. As shown in FIG. 13, various additional implementations are possible.

[0088] Values, Features, and Thresholds: As used herein, the terms "values" and "features" are interchangeable and contemplate raw and analyzed data, whether numerical, time-scaled, graphical, or otherwise. In various implementations, such as those described herein, a value, such as a high frequency count, can be compared to a threshold. Alternatively, or in addition, a ratio of two or more values ​​can be compared to a threshold. The threshold can be a predetermined value. The threshold can be based on statistical information regarding the presence or absence, or likelihood, of subsequent occurrence of delirium, such as information from a population of individuals. In certain embodiments, the threshold can be predetermined for one or more patients. In certain embodiments, the threshold can be consistent for all patients. In certain embodiments, the threshold can be specific to one or more characteristics of the patient, such as current health status, age, sex, race, medical history, other medical conditions, and the like. In certain embodiments, the threshold can be adjusted based on physiological data in the patient's electronic medical record (EMR).

[0089] In certain embodiments, the threshold may be a ratio of high frequency to low frequency. In certain embodiments, the threshold may be a ratio of high frequency over a period of time to low frequency over a period of time. Throughout this disclosure, the ratio is referred to as a ratio of high frequency to low frequency, but it is understood that it may be a ratio from low frequency to high frequency, so long as the format of the ratio is consistent throughout the process. For example, the comparison may be a ratio of high frequency to low frequency, or the inverse low frequency / high frequency ratio.

[0090] One or more features or values ​​may be predetermined. For example, the high frequency range may be The range of low frequencies may be predetermined to be greater than a set value. Similarly, the range of low frequencies may be predetermined to be less than a set value. The set value may be the same for all patients or may vary depending on the characteristics of a particular patient.

[0091] Other features or values ​​of one or more signals may be extracted. For example, the signal-to-noise ratio may be determined for other applications. The quality of the data may be assessed by looking for frequencies of non-physiological electrical activity. Data collection and / or interpretation may be limited to stop if data quality falls below an acceptable level.

[0092] Device Characteristics and Signal Collection: The systems and methods described herein can provide a special-purpose screening device 10, system 1 and method 5, that is simple, convenient and easy to use. In certain embodiments, the systems and methods can utilize simplified electroencephalography (EEG) technology for the end user. The systems and methods can automatically interpret data and provide guidance (advice) to medical personnel regarding monitoring, screening, or subsequent onset of delirium by patients. Traditionally, EEG data is visually inspected by trained neurologists, and there is no automation of the process. In certain embodiments described herein, there can be options for interfacing with standard monitoring equipment, mobile devices, cloud technology, etc. to create an automated process.

[0093] Certain embodiments described herein may be useful in a variety of medical fields, including, but not limited to, intensive care, pre- and post-operative care, geriatric care facilities, nursing homes, emergency rooms, and trauma care. Monitoring, screening, or prediction may include: Patient care can be improved in hospitals or other healthcare settings. Patients can also utilize personal medical devices and monitoring to allow them to remotely monitor their condition when not in a healthcare environment. For example, personal medical devices can monitor patients at home or other locations outside of a healthcare facility and provide delirium monitoring, screening or prediction. Remote sensing and / or analysis systems can interface with systems utilized by healthcare professionals.

[0094] As shown in various embodiments, one or more sensors 12 are configured to communicate with a patient 30. In certain embodiments, the one or more sensors 12 may be, but are not limited to, one or more brain sensors, such as an EEG device, such as one or more EEG leads / electrodes. For purposes of this disclosure, the terms "lead" and "electrode" are used interchangeably. The one or more signals may be EEG signals. The EEG signals may include voltage fluctuations resulting from ionic currents in neurons of the patient's brain. In certain embodiments, there may be multiple sensors. In certain embodiments, there may be two sensors, such as two EEG electrodes. The use of fewer than traditional 16 or 24 electrode EEG systems may result in reduced cost and complexity for predicting, screening, or monitoring delirium. In various embodiments, two or more leads or sensors are used. In certain embodiments, 2, 3, 4, 5, 6, 7, 8, 9, or 10 sensors are used. In additional implementations, 11, 12, 13, 14, or 15 sensors are used. In yet other implementations, 15 or more sensors are used. In various implementations, a minimal number of easily placed EEG leads may be used that are fewer than those demonstrated in the prior art, thereby eliminating and / or reducing the need for skilled EEG technicians and / or sub-specialty neurologists. In various embodiments, at least one ground lead is used, and in alternative implementations, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16 or more ground leads are used.

[0095] In certain embodiments, the one or more sensors 12 may be non-invasive. In certain embodiments, non-invasive electrodes may be placed on the patient's skin. In certain embodiments, there may be a minimum of two skin contact points: a sensor and a ground sensor. There may be passive sensors in that no external current flows through these sensors. In certain embodiments, electrically active sensors may be used. In certain embodiments, the one or more sensors may be an adhesive patch with a printed circuit, or a non-adhesive headset that ties the sensor to the skin. In certain embodiments, the one or more sensors may be placed on the patient's head, for example, on the forehead and / or behind one or more of the patient's ears. In certain embodiments, the one or more sensors may be provided with a minimum spacing between the sensors so that they do not contact each other.

[0096] In certain embodiments, minimally invasive or invasive sensors can be used that can provide one or more signals as an indication of a physiological condition, such as brain activity.

[0097] If necessary, one or more signals may be converted from analog to digital signals. Conversion may be performed by the processing device, at the processing device, or before the one or more signals are received at a separate device. If one or more signals are generated or received as digital signals, conversion may not be necessary.

[0098] The one or more signals can indicate one or more brain functions of the patient. In certain embodiments, the one or more signals can provide information about the brainwave activity of the patient. An electroencephalogram can be measured for the patient. In certain embodiments, an electroencephalogram can be performed by EEG, which can be a recording of the brain's electrical activity from the scalp. In certain embodiments, the recorded waveforms can reflect the electrical activity of the cortex, in certain embodiments, the signal strength of the EEG is small and can be measured in microvolts. Traditionally, there are several frequencies and / or frequency bands that can be detected using EEG. The definition of high and low frequencies can vary depending on various factors, including but not limited to the patient population. In certain embodiments, the definition of high and low frequencies can be consistent across the patient population. In certain embodiments, the low frequency is less than about 7.5 Hz, less than about 7.0 Hz, less than about 6.5 Hz, less than about 5.5 Hz, less than about 5 Hz, less than about 4.5 Hz, less than about 4.0 Hz, less than about 3.5 Hz, or less than about 3.0 Hz. In certain embodiments, the high frequency is about 7.5 Hz or more, about 8.0 Hz or more, about 8.5 Hz or more, about 9.0 Hz or more, about 9.5 Hz or more, about 10.0 Hz or more, about 10.5 Hz or more, about 11.0 Hz or more, about 11.5 Hz or more, about 12.0 Hz or more, about 12.5 Hz or more, about 13.0 Hz or more, or about 14.0 Hz or more.

[0099] In certain embodiments, the one or more signals may be a real-time or near real-time data stream. In certain embodiments, the one or more signals may be measured and / or stored prior to processing and / or analysis.

[0100] Although not required, the systems and methods are described in the general context of software and / or computer program instructions executed by one or more computing devices, which may take the form of a traditional server / desktop / laptop, a mobile device such as a smartphone or tablet, a wearable device, medical equipment, other medical systems, etc. The computing devices may include one or more processors connected to data storage for computer program modules and data. Key technologies include, but are not limited to, multi-industry standards of Microsoft® and Linux® / Unix® based operating systems, databases such as SQL Server®, Oracle®, NOSQL®, DB2®, business analytics / intelligence tools such as SPSS®, Cognos®, SAS®, development tools such as Java®, .NET Framework® (VB.NET®, ASP.NET®, AJAX.NET®, etc.), and other e-commerce products, computer languages, development tools, etc. Such program modules may include computer program instructions, such as routines, programs, objects, components, etc., that are executed by one or more processors to perform particular tasks, utilize data, data structures, and / or implement particular abstract data types. Although the systems, methods, and apparatuses have been described in the foregoing context, the activities and operations described below may also be implemented in hardware.

[0101] Analysis and Diagnosis: In certain embodiments, the presence of delirium can be confirmed by the various devices, systems, and methods described herein. In certain embodiments, the absence of delirium can be confirmed or determined by the systems and methods described herein.

[0102] In certain embodiments, the patient 30 may be diagnosed as currently having delirium. A medical practitioner may wish to monitor the patient's state of delirium. The embodiments described herein may provide an efficient and cost-effective system and method for monitoring the patient's state of delirium. The embodiments described herein may allow a medical practitioner to determine whether the patient's delirium is improving, stable, or worsening. In certain embodiments, a medical practitioner may not be certain whether the patient has delirium. For example, the patient may have some, but not all, of the clinical signs and symptoms associated with delirium. The embodiments described herein allow a medical practitioner to determine whether the patient currently has delirium.

[0103] In certain embodiments, patient 30 may not currently be diagnosed with delirium and may not currently have one or more of the clinical signs and symptoms of delirium. For purposes of this disclosure, "clinical signs and symptoms of delirium" refers to any of the following symptoms or symptoms that are consistent with the DSM-5 criteria for delirium (American Psychiatric Association (2013)). Diagnostic and Statistical Manual of Mental Disorders (5th ed.). Washington, DC). In particular, the clinical signs and symptoms of delirium may be as follows: A. Attention problems (i.e., reduced ability to pay attention, concentrate, or sustain attention, and difficulty shifting attention) (disorientation to the environment) and awareness (loss of orientation to the environment). B. Attention disorders occur over a short period of time (usually hours to days), represent a change from baseline attention and cognition, and tend to fluctuate in severity over the course of a day. C. Further cognitive impairment (e.g., memory deficits, disorientation, language, visuospatial abilities, or is a perceptual disorder). D. Disorders in criteria A and C are due to pre-existing, established, or evolving neurological disorders. It is not well explained by cognitive impairment and does not occur in the setting of severely reduced levels of alertness such as coma. E. History, physical examination, or laboratory findings suggest that the attention disorder is due to another medical condition, substance abuse, or other condition. There is evidence that it may be a direct physiological effect of poisoning or withdrawal, or exposure to a toxin, or that it may have multiple etiologies.

[0104] In certain embodiments, the patient 30 is not considered to have been diagnosed with delirium unless all of criteria A-E are met. In other embodiments, all of criteria A-E must be met before the patient is considered to have been diagnosed with delirium. In certain embodiments, the patient may be considered to have been diagnosed with delirium if criteria A-C are met. In certain embodiments, the patient may be considered to have been diagnosed with delirium if two or more of criteria A-E are met. In certain embodiments, the patient may be considered to have been diagnosed with delirium if three or more of criteria A-E are met. In certain embodiments, the patient may be considered to have been diagnosed with delirium if the patient meets criterion A or criterion C. In certain embodiments, the patient may be considered to have been diagnosed with delirium if the patient meets criterion A or criterion C and at least one of criteria B, D, or E. In certain embodiments, the patient may be considered not to have been diagnosed with delirium if three or fewer of criteria A-E are met.

[0105] For purposes of this disclosure, a patient 30 may not have clinical signs and symptoms of delirium if they have not currently been diagnosed by a medical professional as having delirium. The above symptoms are the current guidelines for the diagnosis of delirium based on the DSM-5. In certain embodiments, new and / or updated versions of these guidelines or other guidelines may be used to determine whether a patient has delirium and may have different clinical signs and symptoms. The device 10, system 1 and method 5 herein can predict that a patient will develop delirium before exhibiting one or all of the clinical signs and symptoms, regardless of the criteria utilized for the diagnosis.

[0106] Certain embodiments may provide systems and methods for predicting the subsequent onset of delirium by a patient when clinical signs and symptoms of delirium are not currently present in the patient.

[0107] In certain embodiments, a prediction can be made regarding the likelihood of a subsequent occurrence of delirium for a patient when the patient has not currently been diagnosed with delirium, hi certain embodiments, the patient may not currently have one or more of the clinical signs and symptoms of delirium at the time of determining the likelihood of a subsequent occurrence of delirium.

[0108] In certain embodiments, an indication of the absence, presence, or likelihood of a subsequent onset of delirium can be output to the patient, which can take a variety of forms including notifications, warnings, visual indications, audio indications, tactile indications, reports, medical record entries, emails, text messages, and combinations thereof.

[0109] The indication of the presence or absence or likelihood of subsequent onset of delirium may be a binary indication, such as, for example, "yes" or "no." By way of example, the output may be "yes" that the patient has delirium or "no" that the patient does not have delirium. The output may also be "yes" that the patient may develop delirium or "no" that the patient is unlikely to develop delirium.

[0110] In certain embodiments, the indication of the presence or absence or likelihood of subsequent onset of delirium may be a non-binary indication. The indication may, for example, indicate a percentage risk, i.e., a 70% chance of subsequent onset of delirium. In certain embodiments, the indication of the presence or absence or likelihood of subsequent onset of delirium may be categorical. The indication may, for example, be that the user has a "high", "medium" or "low" risk, as well as intermediate categories such as "medium-high" or "medium-low". The indication may also be any scale, such as 1-5, 1-10, etc.

[0111] The indication of the likelihood that the patient will subsequently develop delirium can be based on a percentage of likelihood. For example, an indication that the patient is at risk for subsequently developing delirium (indication frequency) can be based on a 50% or greater, 55% or greater, 60% or greater, 65% or greater, 70% or greater, 75% or greater, 80% or greater, 85% or greater, 90% or greater, or 95% or greater likelihood that the patient will subsequently develop delirium.

[0112] The instructions can be either delirium monitoring, screening, or prediction. The instructions can also calculate and / or output a confidence score. For example, the instructions can include an 80% confidence score that the patient will develop delirium in a particular time period.

[0113] It should be noted that while the above description is directed to preferred embodiments of the invention, other variations and modifications (improvements) will be apparent to those skilled in the art and may be made without departing from the spirit or scope of the invention. Furthermore, features described in connection with one embodiment of the invention may be used in conjunction with other embodiments, even if not explicitly stated above.

[0114] Although the present disclosure has been described with reference to preferred embodiments, workers skilled in the art will recognize that changes may be made in form and detail without departing from the spirit and scope of the disclosed devices, systems and methods.

Claims

1. a. a sensor configured to record one or more electroencephalographic signals indicative of at least two but not more than twenty electroencephalographic frequencies; b. a processor; c. at least one module, i. performing a spectral density analysis on EEG frequencies over time to establish a ratio of density of high frequency EEG components to density of low frequency EEG components, where high frequency EEG components are equal to or greater than 7.5 Hz and low frequency EEG components are less than 7.5 Hz; ii. establishing and / or adjusting a threshold for said ratio; iii. comparing a ratio of the density of high frequency components of electroencephalograms to the density of low frequency components of electroencephalograms with said threshold value; iv. outputting data indicative of the presence or absence of onset of delirium or an indication of the likelihood of onset of delirium based on a comparison between the ratio of the density of high frequency components of the electroencephalogram to the density of low frequency components of the electroencephalogram and said threshold value; A module configured to execute 23. A system for screening comprising:

2. The system of claim 1, wherein the module is configured to modify and update the threshold over time.

3. The system described in claim 1, wherein the threshold is set and / or adjusted based on the patient's electronic medical record.

4. 1. A delirium screening and thresholding system comprising: a. a handheld screening device comprising a housing and less than five sensors configured to record one or more electroencephalogram signals indicative of electroencephalogram frequencies; b. at least one module, i. collecting brainwave frequencies over time from less than five sensors; ii. performing a spectral density analysis of the EEG frequencies to establish at least one ratio of density of high frequency components of the EEG to density of low frequency components of the EEG, the high frequency components of the EEG being equal to or greater than 7.5 Hz and the low frequency components of the EEG being less than 7.5 Hz; iii. establishing and / or adjusting a threshold for said ratio; iv. comparing the ratio of the density of high frequency components of the electroencephalogram to the density of low frequency components of the electroencephalogram with the threshold value; v. outputting data indicative of the presence or absence of onset of delirium or encephalopathy or an indication of the likelihood of onset based on a ratio of the density of high frequency components of electroencephalograms to the density of low frequency components of electroencephalograms and said threshold value; A module configured to execute Including, the system.

5. The system described in claim 4, wherein the threshold is established and / or adjusted based on the patient's electronic medical record.

Images