Improved detection and analysis of bioacoustic signals

Abstract

This invention provides an intestinal audio analysis device and method for predicting gastrointestinal dysfunction before any clinical signs or symptoms occur. [Solution] A device and method for predicting the risk and likelihood of postoperative gastrointestinal dysfunction based on regression analysis of multiple spectral events related to bowel sounds, wherein erroneously elevated values ​​of those events due to ambient noise are reduced or eliminated.

Description

Background Art

[0001] Postoperative ileus (POI) is an acute paralysis of the gastrointestinal tract that occurs 2 to 6 days after surgery and causes undesirable side effects such as nausea and vomiting, abdominal pain and distension. This occurs most frequently in gastrointestinal surgery. There is a lack of ability to predict which patients are likely to develop these bowel problems, which results in cases of increased incidence of postoperative complications, hospitalizations, and readmission rates. There is a further lack of bowel audio analysis devices and methods that more accurately provide risk level analysis by limiting and / or removing ambient noise that causes false data points, and the devices and methods are more self - contained.

[0002] Therefore, devices and methods for reducing these problems are still needed, and healthcare providers are better equipped to reduce hospitalizations and readmissions and avoid exacerbation through more accurate nutritional intake introduction, which is achieved through an improved ability to obtain more accurate risk level information related to the likelihood of gastrointestinal dysfunction occurring in patients, for example, postoperatively.

Brief Description of the Drawings

[0003] [Figure 1] Shows an example of the device described herein. [Figure 2] Shows a top view of an example of the device described herein. [Figure 3] Shows an exploded view of the device described herein. [Figure 4] Shows an exploded view of the device described herein. [Figure 5] Shows a bottom view of an example of the device described herein. [Figure 6] Shows a front view of a separated housing of an example of the device described herein. [Figure 7] Shows a side view of a separated housing of an example of the device described herein. [Figure 8]This document shows a rear view of the separated housing of an example of the device described herein. [Figure 9] This shows the inside of the top of the housing of an example of the device described herein, without the user interface. [Figure 10] This specification shows an internal view of the lower part of the housing of an example of the device described herein. [Figure 11] This specification shows a bottom view of the housing of a secondary audio acquisition device, including the housing without a base, as an example of the device described herein. [Figure 12] A side view of an example of the device described herein is shown. [Figure 13] A front view of an example of the device described herein is shown. [Figure 14] This shows a rear view of the device described herein. [Figure 15] This is an illustrative spectrogram illustrating the spectral events contained in recorded abdominal sounds. [Figure 16] This graph illustrates the temporal changes in a specific spectral event (MH4) in patients with and without gastrointestinal dysfunction. [Figure 17] This is a flowchart illustrating an embodiment of a method for predicting gastrointestinal dysfunction. [Figure 18] These are block diagrams of embodiments of device architectures, such as those shown in Figures 1-14, which can process collected patient data to support the prediction and risk assessment of gastrointestinal dysfunction. [Overview of the project]

[0004] This invention relates to the use of an intestinal audio analysis device for predicting gastrointestinal dysfunction before any clinical signs or symptoms occur.

[0005] In a particular embodiment, a gut audio analysis device is provided having a patient interface adapted to be attached to the surface of a patient's abdomen for a certain period of time postoperatively. The device has a diaphragm that vibrates in response to bowel sounds to generate sound pressure waves that amplify the bowel sounds emitted within the patient's abdomen. This allows audio data related to the bowel sounds to be collected through an acoustic chamber that receives the amplified sound, and the amplified sound is transmitted to an audio acquisition device located in or near the acoustic chamber. The gut audio analysis device has an external housing and an internal computer, as well as memory and one or more algorithms for providing instructions to the system to perform various steps required to provide accurate collected data and analysis. One or more algorithms can analyze the collected audio data for false signals due to ambient noise, analyze two or more spectral event values ​​(e.g., MH4) over time, and generate a slope of linear regression analysis for comparison with a predetermined slope threshold for making a binary prediction related to the likelihood and risk of GII, the risk of which is communicated, for example, on a user interface display.

[0006] Other embodiments provide methods for using a bowel audio analysis device. In a particular embodiment, a method is provided for predicting the likelihood and risk level of GII occurring before clinical symptoms appear. Bowel sounds are recorded to generate audio data of the patient's bowel sounds. The audio data is processed to identify certain predicted spectral events (e.g., MH4), and the audio data is analyzed for false signals due to ambient noise. Where necessary, certain embodiments provide false signal mitigation to ensure that a more accurate total number of identified predicted spectral events is obtained for linear regression analysis. The resulting slope of the spectral event values ​​over time is used to correlate the slope with the likelihood of gastrointestinal dysfunction occurring (e.g., postoperatively), and based on the correlation, the user is provided with a likelihood-based risk level.

[0007] Embodiments of the present invention provide improved control over ambient noise, which increases the accuracy of data used to perform correlation and, based on that, predict postoperative GII prior to clinical symptoms. [Modes for carrying out the invention]

[0008] Gastrointestinal dysfunction (GII) is a common concern after surgical procedures. This specification discloses a system and method for predicting GII based on a patient's bowel sounds. As described below, the disclosed system and method identify discrete acoustic spectral events within bowel sounds, which can be used to predict subsequent GII and, in some cases, postoperative GII. These spectral events are good indicators of bowel function and / or dysfunction because the sounds are generated by motility activity in the intestines.

[0009] The following disclosure describes various embodiments. These embodiments are merely illustrative implementations of the invention, and it should be understood that other embodiments are possible. All such other embodiments are intended to be within the scope of this disclosure.

[0010] A particular embodiment comprises a device including a housing, a display, a processor, and at least one audio acquisition device. In a particular embodiment, there are primary and secondary audio acquisition devices. In some embodiments, the primary audio acquisition device can be attached to or inserted into the object.

[0011] Figures 1 and 2 provide examples of device embodiments. Figures 1 and 2 partially show an example of a device comprising a housing 101, at least one button 102 for the user to interact with the device and control its operation, and a user interface 103. The user interface may be a graphical user interface (GUI).

[0012] In certain embodiments, the device comprises a secondary audio acquisition device. In some embodiments, the secondary audio acquisition device may be a microphone or other sound receiving mechanism capable of acquiring ambient noise. In other embodiments, signal subtraction of ambient noise from primary microphone data is performed to achieve true noise cancellation. In some embodiments, the devices described herein can be used in methods for predicting or diagnosing GII. In certain embodiments, the systems and methods described herein may comprise predicting GII before the onset of symptoms. In some cases, the secondary audio acquisition device can eliminate interference associated with ambient noise that may interfere with the acquisition of bowel sounds. Such configurations can result in the surprising and unexpected outcome of more accurately predicting GII. In some cases, GII prediction may be postoperative. In some embodiments, the ambient audio acquisition device further comprises a diaphragm.

[0013] Figures 3 and 4 show exploded views of an example device. Figures 3 and 4 show a device comprising a user interface 103, a detachable top portion 104 of the housing, a computer 105, a lower portion 106 of the housing, at least one fastening means 107, a diaphragm 108, a gasket 109, and an adhesive wafer 110. Figure 5 shows an example of a device having a secondary audio acquisition device 111.

[0014] Figures 8-10 show different views of the top 104 and bottom 106 of the housing. Figure 9 shows an internal view of the top 104 of the housing, and Figures 10-11 show different views of the bottom 106 of the housing. Figures 12-14 show different views of a particular embodiment of the device of the present invention.

[0015] The GUI can be configured to enable a user to interact with the device and collect audio data from a subject. Certain embodiments include a housing. The housing can include an upper portion and an adhesive wafer. The size (width and length) of the adhesive wafer can vary. For example, the wafer can be the same width and length as the bottom surface area of the unit or smaller than the bottom surface area of the unit. The wafer can be of different thicknesses. The wafer is formed to best conform the unit to the patient for reliable monitoring / recording / collection of GII events. The housing can include a separable top and bottom.

[0016] In embodiments where the housing includes a separable top and bottom, the separable top and bottom are secured using attachment means. The attachment means can include screws, fasteners, pins, brackets, pegs, rivets, clips, and the like.

[0017] Embodiments of the system can be used to predict gastrointestinal dysfunction. Embodiments of the system can comprise a data collection device, a patient interface, and a computer 105. The computer 105 or the data collection device can be a separate device that collects and / or stores data retrieved from the patient interface and the computer 105. The patient interface can comprise any device capable of collecting audio data generated within the patient's intestinal tract. In some embodiments, the patient interface comprises a portable (e.g., handheld) digital audio recorder. In such a case, the patient interface can comprise an integrated microphone (not shown) used to capture bowel sounds, and the integrated microphone can be housed within or near an acoustic chamber. The acoustic chamber can be formed, for example, between a diaphragm 108 and a lower portion 106 of the housing. The acoustic chamber significantly increases data collection by amplifying the sounds of the patient's digestive tract relative to ambient noise, increases the ability to reduce ambient noise at the time of audio collection, simplifies audio data collection by reducing the inventory of irrelevant components, and provides the ability for continuous audio data collection in combination with the patient interface, the diaphragm, and an audio data collection device, such as a microphone housed within or near the acoustic chamber of the device of the present invention.

[0018] A patient interface may be a device that can be applied directly to the patient's abdomen for the purpose of detecting bowel sounds. In some embodiments, the patient interface comprises a stethoscope head or is similar in design and function. The stethoscope head may comprise a diaphragm positioned in contact with the patient and vibrating in response to sounds generated within the body. These sounds can be transmitted to a microphone that can transmit signals to a computer 105, which can optionally transmit data to a data acquisition device. In some embodiments, the patient interface can transmit sound via a tube extending between the patient interface and a sound acquisition device, such as a microphone, which may be located in an acoustic chamber. In some embodiments, the patient interface can convert sound waves into digital signals that can be processed by the computer 105. For example, sound pressure waves generated by diaphragm vibration can travel through the lumen of the tube to the microphone. In some embodiments, the patient interface may comprise a microphone or other device that converts sound pressure waves into digital signals that can be processed and analyzed by the system to predict GII. For example, a microphone or other device for converting sound pressure waves into digital signals may be housed in or communicate with an acoustic chamber positioned between the diaphragm 108 and the lower part 106 of the housing. In some embodiments, all or part of the patient interface may be disposable to avoid cross-contamination between patients. The patient interface may be used with a disposable sheath or cover (not shown) that can be discarded after use.

[0019] Audio data collected by computer 105 can be stored in internal memory provided by the computer. In some embodiments, the memory may be separate from computer 105 but communicates with computer 105 for data. In some embodiments, audio data can be stored in the device's non-volatile memory (e.g., flash memory). The data can then be transmitted to computer 105 for processing. In some embodiments, the data and data analysis are transmitted via wires or cables used to physically connect computer 105 to a separate processor or computer, such as a laptop or desktop computer or a handheld or portable device, which has the components necessary to receive and transmit the data and data analysis. In some embodiments, data and data analysis can be wirelessly transmitted from the device to a separate processor or computer via a suitable wireless protocol such as Bluetooth® or Wi-Fi (IEEE 802.11), or via a suitable cellular protocol such as Global System for Mobile Communication (GSM®), General Packet Radio Service (GPRS), Enhanced Data Rates for GSM Evolution (EDGE), Universal Mobile Telecommunications Service (UMTS), High Speed ​​Packet Access (HSPA), Code-Division Multiple Access (CDMA), Evolution-Data Optomized (EV-DO, EVDO, 1xEV-DO), Short Message Service (SMS), and Wi-MAX.

[0020] The separate computer may, in some embodiments, be a desktop computer. However, it should be noted that substantially any computing device capable of receiving and processing the audio data collected by computer 105 can be used. Thus, the separate computer may, alternatively, take the form of a mobile computer such as a notebook computer, tablet computer, smartphone, or handheld computer. In some embodiments, computer 105 can process any data generated from the acoustic sounds received from the patient and output the information via the user interface 103. In some embodiments, the user interface may have a display function and may also be a touchscreen that allows the user to interact with the device and control its operation. For example, the device may be equipped with a digital signal processor and appropriate software / firmware that can be used to analyze the collected audio data.

[0021] In some embodiments, patient sounds can be wirelessly transmitted from the patient interface to the device. In some embodiments, the patient interface may have an adhesive surface that allows the interface to be temporarily attached to the patient's skin, similar to an electrocardiogram (EKG) lead. In certain embodiments, patient data can be transmitted from the patient interface to the computer 105 via a wired connection (via wire or cable) or wirelessly.

[0022] In some embodiments, the patient interface may comprise a device having its own integrated microphone (not shown), and the patient's sounds picked up by the microphone can be electronically transmitted to the device along a wire or cable. In some embodiments, the device may comprise components designed to dock with a patient monitoring system, which may be located with the patient for a continuous period of time, such as several hours, several days, or several weeks. Such patient monitoring systems are currently used to monitor other patient parameters, such as blood pressure and oxygen saturation. For example, the patient monitoring system may comprise a docking station and associated display. In certain embodiments, this allows warnings or notifications to be received at a secondary location, such as a nurse's station. The device may also be designed to dock in a free bay of the station before use. Other means of connecting the device of the present invention include, for example, via various ports known in the industry for these types of electronic systems, including PS / 2, serial ports (e.g., DB-25, DE-9, or RS-232, or COM Port), parallel ports or Centronics 36-pin ports, audio ports, S / PDIF, TOSLINK, video ports, Digital Video Interface (DVI) (e.g., Mini DVI, Micro DVI), DisplayPort, RCA connectors, component video, S-video, HDMI®, USB (e.g., Type A, Type C), RJ-45, RJ-11, e-SATA, and the like. As understood, certain adapters for hard-connecting the present invention to a patient monitoring system are also intended. Types of patient monitoring systems are known and include, for example, standard in-hospital patient monitoring systems for fixed-position systems, as well as mobile and / or portable monitoring systems, remote and / or wireless, home systems, and / or combinations thereof.

[0023] In some embodiments, the device may not have an internal power supply and therefore can only collect patient data when docked. In some embodiments, the device may have a power connection such as an A / C connection or an internal power supply such as a battery. In some embodiments, the device may have a photovoltaic power supply such as a solar cell and / or a thermal power supply. In some embodiments, the patient interface may have a power supply such as a battery, a photovoltaic device, and / or a thermoelectric device that can be used to collect acoustic waves and / or process the acoustic waves into data that can be used to predict GII. As an example, the device may have electrical pins to electrically connect the device to the patient monitoring system for the purpose of receiving power and transferring the collected data to the patient monitoring system. The patient data can be stored in the memory of the patient monitoring system and / or transmitted to a central computer for storage in association with patient records in an associated medical record database.

[0024] The device may be equipped with an electrical port capable of receiving a wire or cable plug. In addition, the device may be equipped with one or more indicators, such as light-emitting diode (LED) indicators, that communicate information to the operator, such as positive electrical connection to the patient monitoring system and patient signal quality.

[0025] The device can be coupled with a patient monitoring system. In some embodiments, instead of an external patient interface, the device may have an internal patient interface designed to collect sound from within the peritoneal cavity. As an example, the patient interface may include a small-diameter microphone catheter that is left in place after the surgery is completed, similar to a drainage catheter. Such a patient interface may be particularly useful when the patient is obese and it is more difficult to obtain a high-quality signal from the surface of the skin. To avoid electric current flowing through the patient, the patient interface may include a laser microphone. In such a case, a laser beam is directed through the catheter and reflected off a target inside the body. The reflected light signal can be received by a receiver that converts the light signal into an audio signal. Tiny differences in the distance the light travels as it reflects off the target are detected by interferometry. In certain embodiments, the patient interface may include a microphone positioned at the tip of the catheter.

[0026] The device may include a central processing unit (CPU) such as computer 105, or other processing devices such as a microprocessor or digital signal processor. The device may include memory which may consist of any one or a combination of volatile memory elements (e.g., RAM) and non-volatile memory elements (e.g., flash, hard disk, ROM).

[0027] The user interface 103 may include components that allow the user to interact with the device. The user interface 103 may include, for example, a keyboard, a mouse (mousepad), an interactive touch display, and / or a display device such as a liquid crystal display (LCD). The device may include one or more buttons 102 that allow the user to operate the user interface 103, such as using the device, navigating menus, viewing data, etc. Alternatively, or in addition, the user interface 103 may include one or more buttons and / or a touchscreen. One or more I / O devices may be adapted to facilitate communication with other devices and may include one or more electrical connectors and a wireless transmitter and / or receiver. In addition, if the device also collects and / or processes data extracted from the patient, the I / O device may include a microphone.

[0028] In some embodiments, the memory may be a computer-readable medium and may store various programs (i.e., logic) including an operating system and a bowel sound analyzer. The operating system may control the execution of other programs and may also provide scheduling, input / output control, file and data management, memory management, and communication control and related services. The bowel sound analyzer may comprise one or more algorithms configured to analyze bowel audio data for the purpose of predicting the likelihood that a patient will develop GII. In some embodiments, the analyzer may perform analysis on correlation data stored in a database (such a database may be stored in computer 105) and present a predictive index of GII risk to a user (e.g., a doctor, hospital staff, and / or clinician). In some embodiments, the analyzer may use target signal parameters, signal-to-noise ratio parameters, and noise power estimation parameters to identify specific spectral events of interest. Then, using decision tree analysis of the number of predicted spectral events in a given time interval, it may communicate high, medium, or low risk of GII. In some embodiments, the risks associated with each risk level are 83%, 30%, and 0%, or approximately thereto, respectively.

[0029] In certain ways of using the device, a patient's bowel sounds can be recorded to generate audio data. As described above, the sounds can be acquired non-invasively, for example, using a stethoscope head or other patient interface placed on or near the patient's abdomen. Alternatively, the sounds can be collected using a device that extends into the patient's peritoneal cavity. The sounds can be recorded early in the postoperative period, for example, on the day of surgery or immediately after surgery, and over any required length of time thereafter. In some embodiments, the acoustic sounds, acquired immediately after surgery, may continue to be monitored and analyzed over several hours, days, or weeks, as will be understood by those skilled in the art in light of this disclosure. In some embodiments, the sounds are collected continuously. In some embodiments, the sounds are collected over a time period that is a predetermined interval. Regardless of when the sounds are recorded, they can be recorded for a duration long enough to allow for the identification of spectral events that predict the occurrence of bowel function and / or dysfunction. As an example, sounds are recorded over a period of approximately 4 to 6 minutes. In some embodiments, all sounds in the range of 20 to 20000 Hz are recorded. In some embodiments, sounds outside the range of human hearing are collected and analyzed. Filters can be applied to reduce the range of frequencies recorded, and thus reduce the amount of data analyzed. In some embodiments, the secondary audio acquisition device 111 can work in conjunction with a patient interface to provide improved analysis results. For example, the secondary audio acquisition device may include a microphone capable of measuring ambient noise in the vicinity of the patient. The ambient noise can then be subtracted from the acoustic sounds collected from the patient to ensure that any acoustic sounds analyzed by the device are actually collected from the patient and not from any external sources. In certain embodiments, for example, a secondary microphone and an integrated microphone unit collect audio data independently in a synchronized manner.Both sets of audio data are processed through an MH4 detector to determine if ambient noise is potentially triggering false MH4 event counts coming from the integrated microphone, and if certain criteria are met by the ambient sound(s), the ambient data count can be subtracted from the final count. In some embodiments, filters can be used so that only sounds with frequencies of approximately 700–1500 Hz are recorded or analyzed, but filters can be incorporated to allow the device to analyze only sounds within a given frequency range. As part of mitigating the effects of ambient noise, a warning indicator or alarm mechanism may be part of the device, which notifies, for example, a physician, caregiver, nurse, or clinician that the ambient noise(s) must be reduced. There are scenarios where the ambient noise level is too high for the device's noise reduction strategy to function. In such scenarios, meaningful MH4 values ​​cannot be obtained. Warning indicators, which may be any combination of light, noise, audible sound, or graphics on the device display, can be used to indicate to the user that the ambient noise needs to be reduced and / or controlled. The threshold for triggering the warning indicator can be based, for example, on the current ambient noise decibel level from the outward-facing microphone, an analysis of the signal / noise ratio from the primary outward-facing microphone, and processing of sound from the outward-facing microphone for false triggering MH4 values ​​that exceed a preset threshold. Although we have described sound as being "recorded," it will be understood that sound can alternatively be simply acquired and processed in real time (as described below) without actually recording the sound.

[0030] Once audio data is generated, the data can be processed, for example, in real time, to identify one or more predictive spectral signals. Real time could mean, for example, that the data is processed while the sound is being acquired through the device as it is being produced by the patient. As described above, sounds produced by the intestines may be the result of peristalsis, and from peristalsis, the device can predict the likelihood of GII, meaning that the device indicates (before GII) the likelihood that GII will occur after data collection and analysis. The sounds collected and / or analyzed by the device can therefore provide an indicator of how the intestines are functioning. For example, paralysis of a significant portion of the intestinal tract proportionally reduces the number of high-energy propulsive contractions in the gastrointestinal tract, which results in some loss of higher-energy, and from this, higher-frequency acoustic spectra that are typical of a normally functioning intestine. As described below, it has been found that certain predetermined spectral events can be identified in the sound that highly predict whether or not GII is likely to occur. As similarly described below, each of these predetermined spectral events is defined by certain characteristics or parameters such as their frequency, amplitude, duration, and temporal separation from other spectral events.

[0031] After spectral events are identified, their number can be summed up over a specified time duration (e.g., the total duration of the recording). The total number of identified predicted spectral events is timestamped and written to a data buffer (buffer) in the device of the present invention. As understood, the buffer may be, for example, in physical memory storage used to temporarily store data, or it may be in fixed memory of the hardware, or it may be a virtual data buffer in software pointing to a location in physical memory. At this point, the total number of spectral events can be compared to correlation data that correlates the number of spectral events with the likelihood of subsequent GII. As an example, a spectral event referred to as "MH4" was identified in the study described below. For MH4, there is a high risk of GII if the number of observed MH4 events is less than approximately 21 during a 4-minute recording, a medium risk of GII if the number of observed MH4 events is more than approximately 21 but less than approximately 131 during a 4-minute recording, and a low risk of GII if the number of observed MH4 events is more than approximately 131 during a 4-minute recording. The number of a given spectral event can therefore be used as an index to convey the magnitude of the risk for GII, with smaller numbers indicating higher risk and larger numbers indicating lower risk.

[0032] If the possibility of a subsequent GII is identified, the risk can be communicated to the user. For example, a computer 105 or other device (such as a separate computer) can be used to perform the analysis and display the risk level on an associated display such as the user interface 103, while the prediction can be displayed on a separate display such as a separate computer monitor. In some embodiments, the prediction can be automatically communicated to the hospital charting and / or recording system so that the prediction is automatically stored along with the patient record. In some cases, the data can be anonymized and used as aggregated data. In some embodiments, the risk can be communicated as an index (i.e., a number). In other embodiments, the risk can be indicated as “high,” “moderate,” or “low.” In any case, appropriate measures can be taken regarding the indication, and appropriate measures may include allowing or prohibiting oral nutrition intake. In particular, further recording and analysis can be performed on the patient in the days following surgery to assess bowel function and to support the initial patient assessment.

[0033] Any of the features described herein can be combined or substituted in such a way as can be understood by those skilled in the art, taking into account the disclosures and descriptions provided herein.

[0034] As can be understood from the methods described above, the risk of GII can be assessed in much the same way that the risk of cardiac problems can be non-invasively assessed using EKG. In some embodiments, the risk assessment can be performed in real time.

[0035] A clinical study was conducted to evaluate the feasibility of the disclosed system and method. One objective of this study was to confirm that spectral events present in bowel sounds early in the postoperative period actually correlate with GII before the subsequent onset of clinical signs and symptoms. Another objective of this study was to develop a model for predicting GII that could be implemented as a simple, non-invasive point-of-care test, enabling hospitals and other facilities to risk stratify patients for the development of clinically significant GII using bowel sound analysis.

[0036] In this study, patients scheduled for hospitalization and surgery were recruited using an IRB-approved protocol. Patients undergoing abdominal and non-abdominal surgery were included. Patients admitted to the ICU postoperatively were excluded from the remainder of the study.

[0037] For the study, a device for digitally recording abdominal sounds was assembled using a dual-channel digital audio recorder (Microtrak II, M-Audio Corp., Irwindale, CA), a condenser microphone (ATR35s, Audio-Technica Ltd, Leeds, UK), a stethoscope tube, and a stethoscope head. To record bowel sounds, the stethoscope head was placed against the upper and lower anterior abdominal wall, and both channels were recorded simultaneously for a period of 5-6 minutes. A standardized tone was also applied to each recording to calibrate the audio levels.

[0038] Bowel sounds were recorded by the research team immediately before surgery and then each day postoperatively. The research team also collected clinical outcome data daily. Variables associated with the development of GII are shown in Table 1. The clinical team providing patient care was not informed of the results of the audio recordings. [Table 1]

[0039] Subsequently, the audio recordings were processed using a digital signal processing algorithm. The algorithm was applied in an iterative manner, focusing on identifying spectral events in the early postoperative period that would predict the onset of GII preoperatively or during the remainder of the hospital stay. Five types of spectral events across different parts of the audible spectrum were ultimately used for analysis. Each type of spectral event was defined by unique target signal parameters (minimum and maximum frequencies, minimum and maximum durations, and minimum separation), signal-to-noise ratio parameters (minimum occupancy, signal-to-noise threshold), and noise power estimation parameters (block size, hop size, percentile). The five spectral events were designated H4, M4, L4, ML4, and MH4, with their parameters shown in Table 2. Spectral events were counted over 4-minute time intervals. GII was defined as the presence of vomiting, the need for nasogastric intubation, or reversal of eating and drinking. [Table 2]

[0040] RavenPro 1.4 software was used for visualization, analysis, and measurement of the recorded audio signals. Statistical analysis was performed using PASW18 and Clementine 10.1.

[0041] Thirty-seven patients were enrolled in the study. Five patients were excluded due to postoperative ICU admission. Two patients who were discharged on the day of surgery were excluded because postoperative data could not be obtained. Of the remaining 30 patients, 11 were male and 19 were female. The mean age was 52 years (SD=12). Five patients underwent external abdominal surgery, and 25 underwent internal abdominal surgery. Subsequently, nine patients (30% of the total) developed Grade II, all within the first four days postoperatively. Of these patients, four started at POD1, one at POD2, and four at POD4.

[0042] Three examples of spectral events are shown in the spectrogram in Figure 15. The mean number of each specified spectral event was calculated for patients who subsequently developed or did not develop GII. A two-tailed t-test was then used to assess the significance of any difference. Spectral events obtained from POD0 did not correlate with the subsequent development of GII. Spectral events obtained from POD1, however, were found to correlate with the subsequent development of GII. Specifically, the MH4 spectral event had a mean count of 154 in patients without subsequent GII and a mean count of 44 in patients who developed GII (p=.004).

[0043] Figure 16 shows an example of GII prediction based on data collection as described herein, and a graph plotting the temporal changes in MH4 spectral events. The results of the study confirmed that spectral events present in bowel sounds early in surgical hospitalization actually correlate with GII before the onset of clinical signs and symptoms. In particular, MH4 was found to be highly and significantly separated by the presence of subsequent GII. Therefore, a predictive model based on MH4 measurement can be used to assess patients as high-risk, medium-risk, and low-risk for GII. Significantly, patients in the low-risk group did not develop GII. In this study, the predictive value for the low-risk classification for no GII was 100%, while the predictive value for the high-risk classification for GII was 83%. 30 percent (30%) of medium-risk patients experienced GII.

[0044] As embodiments of the present invention have been described herein and above, Figure 17 provides a flowchart of an embodiment of the method of the present invention for predicting GII. As illustrated, once the device of the present invention is positioned on the patient, in block 100, the patient's bowel sounds are recorded to generate audio data. As described herein, the sounds can be acquired non-invasively through a patient interface to the patient's skin, on or virtually near the patient's abdomen. The recorded audio data from block 100 is then processed in block 102 to identify predictive spectral events (e.g., MH4). The processed audio data from block 102 is then analyzed in block 104A for false signals due to ambient noise. If the device of the present invention does not indicate that such false signals require mitigation, the total number of identified predictive spectral events is calculated in block 106. However, if the device of the present invention detects a number of false signals exceeding a threshold in block 104A, the device recalculates the device, performing mitigation of false signals due to ambient noise by adjusting the number of identified predictive spectral events to account for the false signals, while alerting / warning the user of the need to reduce or control ambient noise. Once mitigation in block 104B is complete, block 106 provides the total number of identified predicted spectral events, as described above for the non-mitigation determination from block 104A. Using the total number of identified predicted events calculated in block 106, the timestamps and the number of spectral events are written to a buffer in block 108A, and the slope of the spectral event values ​​over time is generated from a regression analysis of the buffered data in block 108B. Then, in block 110, the slope values ​​are compared with data that correlates the slope with the likelihood of future GI occurrences. This embodiment provides a significant improvement over other methods by analyzing the slope of two or more spectral event values ​​(i.e., MH4 values) over time. For example, MH4 values ​​are obtained between approximately 7 and 12 hours postoperatively, and the slope of a linear regression analysis of these points is calculated.Rather than a single MH4 value collected at 12 hours postoperatively acting as a predictor, the calculated slope of the present invention is compared to a predetermined slope threshold to perform binary prediction. While this aspect of the present invention corrects potential misinterpretations resulting from a single MH4 value that may be falsely elevated due to ambient noise interference, the present invention provides analysis based on multiple values ​​and results, along with analytical results that are more resilient to external interference such as ambient noise. This ability to mitigate false signals greatly enhances risk level analysis and strategy, especially when used with the device of the present invention equipped with an acoustic chamber as defined herein. From the correlation in block 110, the device of the present invention shows the user the risk level in block 112 for decision-making regarding preparation, prevention, treatment, and other strategies to address the likelihood of a predicted GII event occurring. As discussed herein, devices and methods are provided that enable the collection, analysis, and correlation of real-time data to provide results predicting that GII will occur at a particular risk level.

[0045] Figure 18 illustrates an exemplary architecture for a device 72 that can be used in a system for predicting gastrointestinal dysfunction to analyze collected patient data. For example, the architecture shown in Figure 18 may be an architecture for the device of the present invention comprising, for example, the computer 105 and user interface 103 of Figures 3-4, and data acquisition devices, processing devices, user interfaces, I / O devices (including microphones), memory (including operating systems, bowel sound analyzers, databases, etc.) as described herein. Furthermore, it should be noted that the exemplary architecture can be distributed across one or more devices.

[0046] As shown in Figure 18, the device 72 generally comprises a processing device 74, memory 76, user interface 78, and input / output devices 80, each of which is coupled to a local interface 82 such as a local bus.

[0047] The processing device 74 may include a central processing unit (CPU) or other processing devices such as a microprocessor or a digital signal processor. The memory 76 includes one or a combination of volatile memory elements (e.g., RAM) and non-volatile memory elements (e.g., flash, hard disk, ROM).

[0048] The user interface 78 comprises components through which the user interacts with the device 72. The user interface 78 may include, for example, a keyboard, a mouse, and a display device such as a liquid crystal display (LCD). Alternatively, or in addition, the user interface 78 may include one or more buttons and / or a touchscreen. One or more I / O devices 80 are adapted to facilitate communication with other devices and may include one or more electrical connectors and wireless transmitters and / or receivers. In addition, if device 72 is a data acquisition device, the I / O devices 80 may include one or more microphones 84.

[0049] Memory 76 is a computer-readable medium that stores various programs (i.e., logic) including the operating system 86 and the bowel sound analyzer 88. The operating system 86 controls the execution of other programs and provides scheduling, input / output control, file and data management, memory management, and communication control and related services. Device 72, and preferably memory 76, includes a data buffer to which timestamped spectral events are written for analysis via one or more algorithms and / or other software for comparing slope values ​​with data for correlation of slope with the likelihood of subsequent GII, from which a risk level is determined. The bowel sound analyzer 88 includes one or more algorithms configured to analyze bowel audio data for the purpose of predicting the likelihood that a patient will develop GII. In some embodiments, the analyzer 88 performs its analysis on correlation data stored in the database 90 and presents a GII risk prediction index to the user (e.g., a doctor or hospital staff). In some embodiments, the analyzer 88 uses target signal parameters, signal-to-noise ratio parameters, and noise power estimation parameters to identify specific spectral events of interest. Next, decision tree analysis of the number of predicted spectral events within a specified time interval can be used to communicate high, medium, or low risk for GII. In some embodiments, the risks associated with each risk level are 83%, 30%, and 0%, or approximately thereto, respectively.

[0050] A robust model is thought to be possible by monitoring bowel sounds over longer time intervals, such as 24-hour periods, from larger datasets of patients. Serial recordings, with data averaging and the addition of additional types of spectral analysis, may improve the predictive accuracy of the disclosed techniques. Future trials are anticipated to focus on collecting larger datasets, validating the proposed predictive models, refining the spectral events analyzed, evaluating alternative timings for data collection, and developing broadly applicable predictive models. In addition, further development of reliable techniques for rapid point-of-care data serial acquisition and analysis would be invaluable in scaling these studies and ultimately in any clinical use. In any case, the studies described above support the feasibility and potential of using acoustic spectral analysis in the study of GII and other gastrointestinal disorders.

[0051] Colic is a leading cause of premature death in horses. While colic refers to abdominal pain, it can reflect several underlying disorders affecting the large intestine, including obstruction, spasm, or torsion. Early diagnosis is key to preventing death, but early diagnosis is difficult in larger mammals such as horses. Since the underlying causes of colic reflect mechanical changes in the intestines, they can be recognized as slight variations in the type of noise produced from regular contractions of the intestines.

[0052] The devices and systems of the present invention can be modified to perform audio spectral analysis of bowel sounds in large mammals such as horses. In another aspect of the present invention, a series of devices can be used to monitor multiple horses simultaneously or at overlapping times, and the devices are networked to a gateway device for transmitting data and / or analysis to at least one other receiving system. The devices and systems are configured to be attached to the horse's body (e.g., torso or abdomen / lower abdomen) (i.e., via straps, adhesive, elastic bands, other similar mechanisms, and / or a combination thereof), and the device microphones will be positioned facing the body. Data collected by the devices can be analyzed by the devices. The resulting analysis can be accessed and reviewed from the devices as described herein, or the data analysis can be transmitted to a gateway device located in an area where, for example, a barn, animal shelter, field, pasture, or essentially an area where the gateway device receives information from the devices attached to the horses and transmits that information to a device in another location so that the caretaker can review the analysis to determine the necessary actions for the horses. The audio spectral analysis algorithm should be specifically tuned for early identification of spectral events that predict the likelihood of colic starting or the initial stages of colic, and should have a power consumption strategy that allows for use for several weeks to several months before the device needs to be replaced. While ambient noise near and around horses differs from ambient noise in hospitals or other places where humans are cared for, it should be understood that the device of the present invention can be easily tuned to account for ambient noise near and around horses. The use of warning indicators regarding notification and / or alarms regarding the need to control ambient noise is described above herein.

[0053] While the foregoing description applies to preferred embodiments of the present invention, it should be noted that other modifications and alterations will be obvious 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 present invention may be used in conjunction with other embodiments, even if not explicitly stated above.

Claims

1. A patient interface adapted to be attached to the surface of the patient's abdomen, A diaphragm that vibrates in response to bowel sounds, creating sound pressure waves that amplify the bowel sounds emitted within the patient's abdomen, and providing audio data for collection. An acoustic chamber adapted to receive sound pressure waves caused by the vibration of a diaphragm that amplifies the aforementioned bowel sounds, An audio acquisition device located in or near the aforementioned acoustic chamber, A housing containing a computer, A memory containing one or more algorithms, A user interface equipped with a display, An intestinal audio analysis device comprising, At least one of the one or more algorithms described above analyzes the collected audio data for false signals caused by ambient noise, At least one of the one or more algorithms described above analyzes two or more spectral event values ​​over time, At least one of the one or more algorithms described above is responsible for generating the slope of the linear regression analysis of the two or more spectral event values ​​over time. The slope of the linear regression analysis is compared to a predetermined slope threshold in order to perform binary prediction in the device.

2. The device according to claim 1, wherein the display comprises a graphical user interface.

3. The device according to claim 1, wherein the device is configured to collect and analyze audio signals from a patient to predict the likelihood of gastrointestinal dysfunction before the clinical signs of such dysfunction occur.

4. The device according to claim 1, wherein the housing comprises an upper part and a base.

5. The device according to claim 1, further comprising a secondary audio acquisition device.

6. The device according to claim 5, wherein the secondary audio acquisition device is configured to acquire ambient noise.

7. The device according to claim 1, wherein the patient interface collects acoustic sound waves, and at least one audio acquisition device collects processed audio data from the acoustic sound waves.

8. The device according to claim 5, wherein the secondary audio acquisition device further comprises a diaphragm.

9. The device according to claim 1, wherein the housing further comprises a detachable top and bottom portion.

10. The device according to claim 9, wherein the detachable top and bottom portions are secured using mounting means.

11. A method for predicting the likelihood of gastrointestinal dysfunction occurring, wherein the method is: Recording bowel sounds to generate audio data from patients, Processing the audio data to identify predicted spectral events, The audio data is analyzed for erroneous signals caused by ambient noise, and if the analysis of the erroneous signals indicates that mitigation is necessary, the erroneous signals are mitigated. Obtain the total number of identified predicted spectral events, Perform a linear regression analysis to obtain the slope of spectral event values ​​over time. Correlating the aforementioned slope with the likelihood of gastrointestinal dysfunction occurring, To provide users with a risk level based on the aforementioned possibilities A method comprising, wherein the method is performed on the device described in claim 1.

12. The method according to claim 11, wherein the mitigation step comprises adjusting the number of identified predicted spectral events by taking into account the false signals caused by ambient noise, resulting in the total number of identified predicted spectral events.

13. The method according to claim 11, wherein the step of performing the above steps comprises providing a timestamp and writing the total number of identified predicted spectral events to a buffer of the device.

14. The method according to claim 13, wherein the linear regression analysis is performed using the data written to the buffer.

15. The method according to claim 11, wherein the correlation step is performed by comparing the value of the slope with data that correlates the slope with the likelihood of gastrointestinal dysfunction occurring later.

16. The method according to claim 11, further comprising the step of warning the user if an analysis of the false signal indicates that mitigation is necessary.

17. The method according to claim 16, wherein the warning is provided with a warning indicator as part of the device, and the warning indicator is selected from a list of light, noise, graphics on a user interface display, or a combination thereof.

18. The aforementioned warning step is, The method according to claim 16, which is activated by 1) a preset ambient noise decibel level from a secondary audio acquisition device, 2) analysis of the signal-to-noise ratio from a primary audio acquisition device versus a secondary audio acquisition device, and 3) processing of noise received by the secondary audio acquisition device for false triggering of spectral events exceeding a preset threshold.

19. The method according to claim 11, wherein the predicted spectral event is MH4.

20. The device according to claim 1, further comprising a warning indicator for false signals caused by excessive ambient noise.

Images