Diagnostic method and system using sound detection

Abstract

A method of monitoring a medical treatment device includes detecting a sound of the medical treatment device with a detection device, determining whether the sound is an irregular sound, and deploying an intervention upon determining that the sound is an irregular sound.

Description

Technical Field

[0001] The present disclosure relates to diagnostic methods and systems using sound detection. Background Art

[0002] Dialysis is a treatment used to support patients with insufficient kidney function. The two main types of dialysis are hemodialysis and peritoneal dialysis.

[0003] During hemodialysis ("HD"), a patient's blood is passed through the dialyzer of a dialysis machine while a dialysis solution or dialysate is also passed through the dialyzer. A semipermeable membrane in the dialyzer separates the blood within the dialyzer from the dialysate and allows diffusion and osmotic exchanges to occur between the dialysate and the bloodstream. These exchanges across the membrane allow waste products, including solutes like urea and creatinine, to be removed from the blood. These exchanges also regulate the levels of other substances in the blood, such as sodium and water. In this way, the dialysis machine acts as an artificial kidney for purifying the blood.

[0004] During peritoneal dialysis ("PD"), the patient's peritoneal cavity is periodically infused with a dialysis solution, or dialysate. The membrane lining of the patient's peritoneum acts as a natural semipermeable membrane that allows diffusion and osmotic exchange between the solution and the bloodstream. These exchanges across the patient's peritoneum, like the continuous exchanges across the dialyzer in HD, allow waste products, including solutes like urea and creatinine, to be removed from the blood, and regulate the levels of other substances in the blood, such as sodium and water.

[0005] Many PD machines are designed to automatically fill, dwell, and drain dialysate into a patient's peritoneal cavity. Treatments typically last for many hours and usually begin with an initial drain cycle to empty the peritoneal cavity of used or spent dialysate. The sequence then proceeds through a succession of fill, dwell, and drain phases, one after the other. Each phase is called a cycle. Summary of the invention

[0006] Certain methods and systems described herein enable diagnosis of mechanical problems upon detection of unique sounds occurring around a patient treatment. Sounds occurring before, during, or after treatment can provide insight into machine conditions or treatment events, which can assist or trigger certain needed clinician or technical support through automated, real-time intervention. For example, home treatments can particularly benefit from remote and automated insights into patient treatments being performed, e.g., given that patients often have limited experience with dialysis equipment.

[0007] A range of diagnostics can be derived from audible sounds in the vicinity of a patient undergoing treatment. Each sound (e.g., fluid flow, pump operation, electromechanical components, housing vibrations, machine-generated alarms, etc.) has a unique signature that can be monitored to drive interventions by a defined set of patients, clinicians, and medical device manufacturers. Each sound contains a unique acoustic fingerprint or signature that can be classified based on learned treatment events that represent normal and abnormal conditions to drive specific interventions to attempt to correct the identified problem.

[0008] In one aspect, a method of monitoring a medical treatment device includes: detecting sounds of the medical treatment device with a detection device; determining whether the sounds are irregular sounds; and deploying intervention when the sounds are determined to be irregular sounds.

[0009] In some embodiments, the method includes deploying the intervention including displaying a notification.

[0010] In some embodiments, the notification is a push notification on a smartphone.

[0011] In some embodiments, deploying the intervention includes initiating an audio prompt on a device associated with a patient of the medical treatment device.

[0012] In some embodiments, deploying the intervention includes initiating a telephone call to a telephone associated with a patient of the medical treatment device.

[0013] In some embodiments, deploying the intervention includes initiating an audio prompt on a device associated with a patient and / or clinician of the medical treatment device to restart the medical treatment device.

[0014] In some embodiments, deploying the intervention includes initiating the audio prompt on the device associated with the patient and / or clinician to restart the medical treatment device after confirming that the medical treatment device is in a pre-treatment stage.

[0015] In some embodiments, the irregular sound is associated with a pump of the medical treatment device.

[0016] In some embodiments, determining that the sound is the irregular sound comprises comparing the sound to sound samples stored in a database.

[0017] In some embodiments, the sound samples are uploaded to the database and recognized through machine learning.

[0018] In some embodiments, the irregular sound indicates a malfunction of the medical treatment device.

[0019] In some embodiments, the irregular sound indicates a disconnected vein.

[0020] In some embodiments, the irregular sound indicates a pump malfunction.

[0021] In some embodiments, the medical treatment device is a hemodialysis machine.

[0022] In some embodiments, the medical treatment device is a peritoneal dialysis machine.

[0023] In one aspect, a medical treatment system includes a medical treatment device, a detection device configured to detect sounds of the medical treatment device, and a processor configured to determine whether the sounds are irregular sounds and deploy intervention when the sounds are determined to be irregular sounds.

[0024] In some embodiments, the intervention includes initiating an audio prompt on a device associated with a patient of the medical treatment device to restart the medical treatment device.

[0025] In some embodiments, the detection device comprises a microphone.

[0026] In some embodiments, the processor is configured to determine whether the sound is an irregular sound by comparing the sound with a sound database.

[0027] In some embodiments, the sound indicates a malfunction of the medical treatment device.

[0028] In some embodiments, the sound indicates a vein disconnection.

[0029] In some embodiments, the sound indicates a pump malfunction.

[0030] In some embodiments, the sound is rattling of a housing of the medical treatment device. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 is a schematic illustration of a hemodialysis system including a hemodialysis machine in communication with a server and a detection device.

[0032] Figure 2A and Figure 2B Respectively in Figure 1 Illustration of the proper connection and venous disconnection of arterial and venous patient lines to the patient that occurs during treatment performed by a hemodialysis system.

[0033] Figure 3 yes Figure 1 Illustration of the blood pump of a hemodialysis machine.

[0034] Figure 4 is a schematic illustration of a peritoneal dialysis ("PD") system including a PD cycler in communication with a server and a detection device.

[0035] Figure 5A yes Figure 4 1. A perspective view of a PD cycler and a PD cassette of a PD system of FIG. The door of the PD cycler is in an open position to show the inner surface of the PD cycler engaged with the PD cassette during use.

[0036] FIG. 5B to FIG. 5D yes Figure 4 Schematic cross-sectional views of a PD cassette in a cassette compartment of a PD cycler of a PD system of FIG. 1 during different stages of operation.

[0037] Figure 6 is a schematic illustration of a hemodialysis system comprising a hemodialysis machine with an integrated detection device, wherein the hemodialysis machine communicates with a server.

[0038] Figure 7 is a method of deploying intervention in response to the detection of sound.

[0039] Figure 8 It is a method to store sound features in a database. DETAILED DESCRIPTION

[0040] The present disclosure generally relates to methods for detecting sound signatures (e.g., the sound of fluid flowing, the whirring of a pump, the rattling of a housing, the sound of an electromechanical component, etc.) and deploying interventions in response to detecting the sound signatures (e.g., instructing a patient to run a diagnostic procedure, sending an alert to a medical provider staff member, prompting an incoming call, sending an app notification, etc.). In some cases, a medical fluid pumping system (e.g., a dialysis system, a peritoneal dialysis system, a hemofiltration system, etc.) includes a detection device (e.g., a microphone, a smart home device, a smartphone, etc.) that actively listens for the sound signature. The detection device actively listens for the sound signature throughout the treatment process. In response to detecting the sound signature, the detection device communicates with a server or with the medical fluid pumping system to automatically perform real-time intervention or trigger support (e.g., notify a clinician, notify a technical support member, run a diagnostic procedure, etc.).

[0041] refer to Figure 1 , the hemodialysis system 100 includes a hemodialysis machine 102. The hemodialysis machine 102 includes a touch screen 104 and a control panel 106. The touch screen 104 and the control panel 106 allow an operator to input various information to the hemodialysis machine 102 or control the hemodialysis machine 102. In addition, the touch screen 104 serves as a display that conveys information to the operator of the hemodialysis system 100.

[0042] The hemodialysis machine 102 includes a processor 108 located inside the machine, which is connected to the touch screen 104 and the control panel 106. The processor 108 is configured to receive data input via the touch screen 104 and the control panel 106, and control the hemodialysis machine 102 based on the received data. For example, the processor 108 can adjust the operating parameters of the hemodialysis machine 102.

[0043] The hemodialysis machine 102 also includes a blood pressure monitor (BPM) 110 for monitoring the patient's blood pressure. The BPM 110 may be an automatic, non-invasive blood pressure monitor that operates according to the blood pressure measurement principle, such as a blood pressure cuff or sleeve. The BPM 110 may measure systolic pressure, diastolic pressure, mean arterial pressure (MAP), and information related to the patient's pulse. Figure 1 The BPM 110 is shown connected to the hemodialysis machine 102 via wires, but in some embodiments, the BPM 110 is wirelessly connected to the hemodialysis machine 102 and is capable of wireless communication with the hemodialysis machine 102 .

[0044] A disposable blood component set 112 forming a blood circuit is connected to the hemodialysis machine 102. During hemodialysis, the arterial and venous patient lines 114, 116 of the blood component set 112 are connected to the patient, and blood circulates through the various blood lines and components of the blood component set 112, including the dialyzer 118. At the same time, dialysate circulates through a dialysate circuit formed by the dialyzer 118 and various other dialysate components and dialysate lines connected to the hemodialysis machine 102. Many of these dialysate components and dialysate lines are located inside the housing of the hemodialysis machine 102, and thus are located in the hemodialysis machine 102. Figure 1 The dialysate passes through the dialyzer 118 together with the blood. The blood and dialysate passing through the dialyzer 118 are separated from each other by a semi-permeable structure (e.g., a semi-permeable membrane and / or semi-permeable microtubes) of the dialyzer 118. Due to this arrangement, toxins are removed from the patient's blood and collected in the dialysate. The filtered blood leaving the dialyzer 118 is returned to the patient. The dialysate leaving the dialyzer 118 includes toxins removed from the blood and is generally referred to as "spent dialysate." The spent dialysate is directed from the dialyzer 118 to a drain system.

[0045] One of the components of the blood component set 112 is an air release device 120. The air release device 120 includes a venting assembly that allows air to pass through while blocking (e.g., preventing) the passage of liquid. Thus, if the blood passing through the blood circuit during treatment contains air, the air will be exhausted to the atmosphere as the blood passes through the air release device 120.

[0046] The dialysate container 122 is connected to the hemodialysis machine 102 via a dialysate supply line 124. A drain line 126 and an ultrafiltration line 128 also extend from the hemodialysis machine 102. The drain line 126 and the ultrafiltration line 128 can be connected to a drain system. The dialysate supply line 124, the drain line 126, and the ultrafiltration line 128 are fluidly connected to various dialysate components and dialysate lines that form a part of a dialysate circuit inside the housing of the hemodialysis machine 102. During hemodialysis, the dialysate supply line 124 delivers fresh dialysate from the dialysate container 122 to the part of the dialysate circuit located inside the hemodialysis machine 102. As described above, the fresh dialysate circulates through various dialysate lines and dialysate components that form a dialysate circuit, including the dialyzer 118. When the dialysate passes through the dialyzer 118, it collects toxins from the patient's blood. The resulting spent dialysate is conveyed from the dialysate circuit to a drain system via drain line 126. When ultrafiltration is performed during treatment, a combination of spent dialysate and excess fluid drawn from the patient is conveyed via ultrafiltration line 128 to a drain system.

[0047] The blood component set 112 is fixed to a module 130 on the front of the hemodialysis machine 102. The module 130 includes a blood pump 132 that can drive blood through the blood circuit. The module 130 also includes various other instruments that can monitor the blood flowing through the blood circuit. The module 130 includes doors, such as Figure 1 As shown, when closed, the door cooperates with the front face of the module 130 to form a compartment sized and shaped to receive the blood component set 112. In the closed position, the door presses certain blood components of the blood component set 112 against corresponding instruments exposed on the front face of the module 130. This arrangement facilitates controlling blood flow through the blood circuit and monitoring blood flowing through the blood circuit.

[0048] A drug pump 134 also extends from the front of the hemodialysis machine 102. The drug pump 134 is a syringe pump that includes a clamping mechanism configured to hold a syringe 136 of the blood component set 112. The drug pump 134 also includes a stepper motor configured to move a plunger of the syringe 136 along the axis of the syringe 136. The shaft of the stepper motor is fixed to the plunger in such a way that when the stepper motor is operated in a first direction, the shaft forces the plunger into the syringe, and when operated in a second direction, the shaft pulls the plunger out of the syringe 136. Thus, the drug pump 134 can be used to inject a liquid drug (e.g., heparin) from the syringe 136 into the blood circuit via the drug delivery line 138 during use, or to draw a liquid from the blood circuit into the syringe 136 via the drug delivery line 138 during use.

[0049] The dialysate circuit is formed by a plurality of dialysate components and dialysate lines located inside the housing of the hemodialysis machine 102, and a dialyzer 118, a dialyzer input line, and a dialyzer output line located outside the housing of the hemodialysis machine 102. The dialyzer input line includes a connector adapted to be connected to one end region of the dialyzer 118, and the dialyzer output line includes a connector adapted to be connected to the other end region of the dialyzer 118.

[0050] In addition to the blood lines forming the main blood circuit described above, a saline delivery line 140 and a drug delivery line 138 are connected to the blood circuit for introducing saline and drugs (e.g., heparin) into the blood circuit. The saline delivery line 138 is connected to a saline bag 142. The drug delivery line 138 is connected to a syringe 136 containing a drug.

[0051] The hemodialysis system 100 also includes a detection device 150, which includes a processor 152, a microphone 154, and a speaker 156. In this example, the detection device 150 is a smart home device that can detect sound and process commands. Some examples of smart home devices include Google Nest, Amazon Echo, and Apple HomePod, but any of a variety of other smart home devices can be used. The processor 152 provides advanced sound processing capabilities so that the detection device 150 can actively monitor unique sound characteristics before, during, or after using the hemodialysis machine 102. For example, active monitoring can include processing the sound detected by the microphone 154 to determine whether the detected sound matches a predetermined sound characteristic.

[0052] The detection device 150 is configured to be able to communicate with the server 160, for example, via an Internet connection. Many medical facilities of various types can access the server 160. In some embodiments, the server 160 can be accessed by most or all medical facilities affiliated with a specific dialysis service provider. In certain embodiments, the server 160 can be accessed by most or all medical facilities in a specific country or multiple countries. The server 160 is filled with data that can be accessed by the detection device 150, as explained in more detail below. The detection device 150 is configured to be able to receive data from the server 160 and actively monitor the hemodialysis machine 102 based on the received data. For example, the server 160 can provide a sound signature that starts the detection device 150 and deploys intervention. The detection device 150 is also configured to be able to upload data to the server 160. As will be discussed below, the detection device 150 can upload the sounds that occur during treatment to the server 160. The server 160 can use the uploaded sounds to determine regular sounds and irregular sounds.

[0053] Figure 2A and Figure 2B The proper connection 200 of the arterial and venous patient lines 114, 116 to the hemodialysis patient and the venous disconnect 220 where the venous patient line 116 has been disconnected or partially disconnected from the patient are shown, respectively. Figure 2A In the illustrated example of a proper connection 200 to a patient shown in FIG, arterial and venous patient lines 114, 116 are connected to an arteriovenous fistula 202, which includes one or more arteries directly connected to one or more veins or venous spaces. An arteriovenous fistula or graft may be used for dialysis to provide increased blood flow. As shown in the proper connection 200, the arterial and venous patient lines 114, 116 are properly positioned and fully inserted into the arteriovenous fistula 202.

[0054] In contrast, Figure 2B In the illustrated venous disconnection 220 shown in , the venous patient line 116 is improperly positioned and thus not fully inserted into the arteriovenous fistula 202. The venous disconnection 220 can cause the patient to lose a large amount of blood within a few minutes. Therefore, it can be advantageous to identify the venous disconnection as quickly as possible.

[0055] The venous disconnection may produce a unique sound that is not usually present during hemodialysis treatment. For example, during normal hemodialysis treatment, the patient or clinician will not usually hear the sound of blood flowing. However, in the case of a venous disconnection or leakage in the fluid lines (e.g., arterial and venous patient lines 14, 116, blood component group 12) of a hemodialysis machine, sounds such as the whooshing sound of blood, liquid flow, or air flow may occur. These irregular sounds can be identified by a detection device, which can be deployed accordingly to intervene, as described below.

[0056] The server 160 contains a list of predetermined sound features and interventions. A sound feature is one or more discrete data elements (intensity value, centroid value, diffusion value, skew value, kurtosis value, reduction value, Mel-frequency cepstral coefficients, etc.) that represent a unique sound. Each predetermined sound feature can be paired with an intervention, and the server 160 can provide the list of sound features and their corresponding interventions to the detection device 150. One of the predetermined sound features can be a discrete data element representing the whooshing sound of blood.

[0057] The detection device 150 actively monitors the hemodialysis treatment. When a venous disconnection occurs during the treatment, it produces a sound of blood swishing (e.g., in the patient line). The detection device processes the detected sound of blood swishing into a sound feature (e.g., one or more discrete data elements). The processor of the detection device then compares the detected sound feature with a list of predetermined sound features provided by the server 160. In this case, the sound feature of the detected blood swishing matches the predetermined sound feature provided by the server 160.

[0058] When the detection device 150 detects a sound signature that matches a predetermined sound signature provided by the server 160, the detection device 150 can deploy an intervention. In the example of a vein disconnection, the detection device 150 can initiate an audio prompt to the patient, a caregiver, or a clinician using the speaker 154. For example, the audio prompt can include instructions to the clinician to occlude blood flow or seek emergency medical care. The audio prompt can immediately draw attention to a vein disconnection that might otherwise go unnoticed for a period of time.

[0059] Using a smart home device as the detection device 150 has several advantages. For example, certain embodiments are advantageous because safety can be increased without modifying the dialysis machine. A patient can use an off-the-shelf device to improve his or her dialysis experience. A patient also does not need to obtain a new dialysis machine to benefit from the present methods and systems. In addition, many smart home devices can be programmed to detect various sound characteristics and include processors that can initiate intervention. Using a smartphone as the detection device 150 provides similar advantages.

[0060] Other sounds may also cause the detection device 150 to initiate intervention. For example, irregular sounds emitted by the blood pump 132 may be an indication of a malfunction. Figure 3 1 is an enlarged view of the blood pump 132 of the hemodialysis machine 102. The blood pump 132 is a peristaltic pump including a roller 302, which compresses the fluid line 304 when the blood pump 132 rotates around the center 306 of the blood pump. The pump 132 usually rotates at a constant speed. During normal operation, the pump 132 does not produce a large noise, such as a banging or running sound. However, during a mechanical failure or malfunction, the roller 302 may hit the wall of the pump 132, or the pump 132 may emit a running sound greater than that of a normally operated pump. In addition, the pump 132 usually produces a constant or continuous noise, and will not usually produce an irregular pattern of sound when operating normally. Banging, running sounds or other irregular noises may indicate that the pump is malfunctioning.

[0061] As described above, the detection device 150 can actively monitor the hemodialysis treatment. When the blood pump 132 fails, it will generate operating sounds. The detection device processes the detected operating sounds into sound features (e.g., one or more discrete data elements). The processor of the detection device then compares the detected sound features with predetermined sound features provided by the server 160. In this case, the server 160 provides a predetermined sound feature representing the pump operating sound.

[0062] When the detection device 150 detects a sound signature that matches a predetermined sound signature associated with a failure of the blood pump 132 provided by the server 160, the detection device 150 can deploy intervention. In the example of the pump running sound, the detection device 150 can use the speaker 154 to initiate an audio prompt to the patient, the escort, or the clinician to request more information from the patient or clinician to identify whether the system is in pre-treatment (e.g., set for treatment), treatment, or after treatment. If the system is in pre-treatment or after treatment, the detection device 150 can initiate an audio prompt to the patient, the escort, or the clinician to restart the system or run a diagnostic program. If the system is in treatment, the detection device 150 can initiate an audio prompt to the patient, the escort, or the clinician to run a diagnostic program after the treatment is completed. If the sound signature does not indicate a serious or catastrophic failure, the patient, the escort, or the clinician can be prompted to complete the treatment. This may be advantageous because early cessation of treatment may cause the patient to retain additional fluid, which may cause swelling, and may cause the patient to retain more waste that should be eliminated. By determining the stage of treatment and the nature of the failure before providing instructions to the patient, caregiver, or clinician, premature discontinuation of treatment can be avoided.

[0063] The detection device 150 and the server 160 can also be used together with other types of medical equipment. Figure 4 As shown, an example of another type of medical device with which the detection device 150 and the server 160 can be used is a PD system 400, which includes a PD cycler 402, a cart 404, four PD solution bags 406 and their associated fluid lines, and a disposable PD box 500 (located at the back of the PD box) for directing fluids during treatment. Figure 4 Behind the door 408 of the PD cycler 402 and Figure 5A 4. The PD cycler 402 includes a touch screen 410 forming a control panel of a patient interface for operation by the patient. The cycler 402 is placed on top of a cart 404 designed to accommodate PD solution bags and associated fluid lines.

[0064] As shown, PD solution bag 406 is suspended on the finger of the side of cart 404. Heating bag 412 is shown as being located in a shallow depression forming heater tray 414, the size and shape of which are designed to accommodate a typical 5-liter PD solution bag. Heater tray 414 has a plurality of heating coils embedded below the surface. Temperature sensors are positioned in the surface of heater tray 414 to track the temperature of the solution in the heating bag for a thermostatic control circuit that opens and closes the heating coils as needed to maintain the PD solution at a desired temperature. When door 408 is closed and securely latched, the box is inserted into the box compartment formed between hinged door 408 and the box interface of PD cycler 402. Because dialysis solution bag 406, heating bag 412, patient line 416 and discharge line 418 are connected to the box, dialysis solution is allowed to flow into and out of the box during use.

[0065] A typical PD machine uses six fluid handling sequences: flush, prime, drain, fill, pause, and dwell. The purpose of the flush sequence is to remove air from all lines (except the patient line 416) and the cassette. This is accomplished by pumping dialysis solution through the lines to be flushed. The prime sequence removes air from the patient line 416 by pumping dialysis solution from the heater bag through the patient line 416.

[0066] During PD treatment, patient line 416 is connected to the patient's abdomen via a catheter. PD treatment usually begins by emptying the patient's abdomen of the spent dialysis solution remaining from the previous treatment. To this end, the PD cycler is activated to draw the spent dialysis solution from the patient into the cassette. The spent dialysis solution is then pumped from the cassette to drain line 418.

[0067] After the spent dialysis solution is drained from the patient, the heated dialysis solution is transferred from the heating bag 412 to the patient. To this end, the PD cycler is activated so that the heated dialysis solution is drawn from the heating bag 412 into the cassette. The heated dialysis solution is then pumped from the cassette to the patient via the patient line 416.

[0068] Once the dialysis solution has been pumped from the heating bag 412 to the patient, the dialysis solution is allowed to stay in the patient for a period of time. During this stay, toxins enter the dialysis solution from the patient's blood through the patient's peritoneum. While the dialysis solution stays in the patient's body, the PD cycler 402 prepares fresh dialysate to be delivered to the patient in a subsequent cycle. In particular, the PD cycler 402 pumps fresh dialysis solution from one of the full dialysis solution bags 406 into the heating bag 412 for heating. To this end, the pump of the PD cycler 402 is activated to draw the dialysis solution from the selected dialysis solution bag 406 into the box. The dialysis solution is then pumped from the box to the heating bag 412.

[0069] After the dialysis solution has been in the patient for the desired period of time, the spent dialysis solution is pumped from the patient to the drain line 418. The heated dialysis solution is then pumped from the heating bag 412 to the patient, where it resides for the desired period of time. These steps are repeated with dialysis solution from two of the three remaining dialysis solution bags 406. The dialysis solution from the last dialysis solution bag 406 is typically delivered to the patient and remains in the patient until a subsequent PD treatment. Once the patient has been filled with dialysis solution, a pause sequence allows the patient to be disconnected from the machine.

[0070] exist Figure 4 In the example shown, the detection device 150 can actively listen for sound signatures before, during, or after using the PD cycler 402. The detection device 150 is also in communication with a server 160. The detection device 150 is configured to receive data from the server 160 and actively listen to the PD cycler 402 based on the received data. For example, the server 160 can provide sound signatures that activate the detection device 150 and deploy interventions. The detection device 150 is also configured to upload data to the server 160. For example, the detection device 150 can upload sounds to the server 160. As will be discussed below, the detection device 150 can upload sounds that occur during treatment to the server, and the server can use the uploaded sounds to determine regular and irregular sounds.

[0071] Figure 5A is a partially exploded view of the PD cycler 402 and the PD cassette 500. The cassette 500 includes a fluid path that allows fluid to flow through the cassette 500. The fluid path in the cassette 500 leads from the pumping chambers 502A, 502B to a connector positioned along the bottom edge of the cassette 500. The connector is configured to receive fittings on the ends of the PD solution bag line, the heating bag line, the patient line 416, and the drain line 418. These fittings can be, for example, double convex fittings. One end of the fitting can be inserted and coupled to its corresponding line, and the other end can be inserted and coupled to its associated connector. By allowing the PD solution bag line, the heating bag line, the patient line 416, and the drain line 418 to be connected to the cassette, the connector allows the PD solution to flow into and out of the cassette 500 during use.

[0072] When the door 408 is in the open position, as Figure 5AAs shown, the cartridge 500 can be loaded onto the cartridge interface 504 by positioning the top portion of the cartridge 500 under the locating pins and pushing the bottom portion of the cartridge 500 toward the cartridge interface 504. The cartridge 500 is sized to remain securely positioned between the locating pins and a lower flange extending from the cartridge interface 504 to allow the door 408 to close over the cartridge 500. When the cartridge 500 is positioned in the cartridge housing 508 between the closed door 408 and the cartridge interface 504, the locating pins help ensure that the pump chambers 502A, 502B of the cartridge 500 are aligned with the piston heads 506A, 506B of the piston pump of the PD cycler 402.

[0073] refer to Figure 5B , with the cassette 500 positioned adjacent to the cassette interface 504, the door 408 is closed over the cassette 500 such that the cassette 500 is substantially contained between the door 408 and the cassette interface 504 within the cassette housing 508. As shown, the hemispherical protrusion 510 of the cassette 500 fits within the recess 512 in the door 408. The PD cycler utilizes a vacuum system to move the portion of the membrane 514 covering the pump chambers 502A, 502B. Therefore, a substantially airtight seal is typically required between the door 408 and the cassette interface 504.

[0074] like Figure 5C As shown, with the cartridge 500 secured within the housing 508, the piston heads 506A, 506B are moved outward (e.g., to a substantially fully extended position) to contact an area of ​​the cartridge membrane 514. In the fully extended position, the inner surface of the membrane 514 is in contact or nearly in contact with the inner surface of the hemispherical protrusion 510 of the rigid base 516 of the cartridge 500. The vacuum system secures the piston heads 506A, 506B to the membrane 514.

[0075] In the case where the piston heads 506A, 506B are fixed to the membrane 514, the Figure 5D As shown, the PD solution is drawn into the pump chambers 502A, 502B of the cartridge 500 by retracting the membrane 514 together with the piston heads 506A, 506B to increase the volume of the pump chambers 502A, 502B. Then, the PD solution is drawn into the pump chambers 502A, 502B of the cartridge 500 by retracting the membrane 514 together with the piston heads 506A, 506B to increase the volume of the pump chambers 502A, 502B. Figure 5C The position shown causes the membrane 514 to move toward the rigid base 516, thereby reducing the volume of the pump chambers 502A, 502B, and fluid can be forced out of the pump chambers 502A, 502B. While forcing the PD solution into and out of the pump chambers 502A, 502B, certain expandable members of the PD cycler 402 can be selectively expanded to direct the pumped PD solution along a desired path in the cassette 500.

[0076] During normal operation, the piston generally does not make irregular noises, such as knocking, and generally has a consistent rhythm. Knocking, irregular rhythms, or other irregular noises may indicate a fault. The detection device 150 can actively monitor, for example, the sounds produced by the operation of the piston pump of the PD cycler 402.

[0077] As described above, the detection device 150 can actively monitor PD treatment. When piston pumps malfunction, they produce irregular rhythms. The detection device processes the sound of the detected irregular sounds and converts the sounds into one or more discrete data elements (i.e., sound features). The processor of the detection device 150 then compares the detected sound features with predetermined sound features provided by the server 160. In this case, for example, the server 160 can provide predetermined sound features representing the impact sound of the piston or the irregular rhythm of the piston.

[0078] When the detection device 150 detects a sound feature that matches a predetermined sound feature provided by the server 160, the detection device 150 can deploy intervention. In the example of the impact sound of the piston or the irregular rhythm of the piston, the detection device 150 can use the speaker 154 to initiate an audio prompt to the patient, the escort or the clinician to request more information from the patient or the clinician to identify whether the system is in pre-treatment (e.g., settings for treatment), treatment or after treatment. If the system is in pre-treatment or after treatment, the detection device 150 can initiate an audio prompt to the patient, the escort or the clinician to restart the system or run a diagnosis. If the system is in treatment, the detection device 150 can initiate an audio prompt to the patient, the escort or the clinician to run a diagnostic program after the treatment is completed. If the sound feature does not indicate a serious or catastrophic failure, the patient, the escort or the clinician can be prompted to complete the treatment.

[0079] In certain embodiments, detection device can directly communicate with hemodialysis machine.For example, one or more hard-wire connections (for example, serial cables, parallel cables, etc.) can be implemented.Wireless technologies such as infrared (IR), radio frequency (RF), Bluetooth, wireless Ethernet or other electromagnetic link technologies can also be used.One or more protocols, transmission standards and data formats can also be used to transmit information between detection device 150 and hemodialysis machine 102.

[0080] In some embodiments, the detection device may not be a separate device, but a part of the medical fluid pumping machine. Figure 6A system 600 including a detection device 650 integrated with a hemodialysis machine 602 is shown. In addition to the integration of the detection device 650 with the hemodialysis machine 602, the components of the hemodialysis machine 602 can be the same as the components of the hemodialysis machine 102. It can be useful to integrate the detection device 650 with the hemodialysis machine 602 because the detection device can communicate directly with the hemodialysis machine. For example, the server 160 can provide a sound feature that activates the detection device 650 and deploys intervention. The detection device 650 is also configured to be able to upload data to the server 160. For example, the detection device 650 can upload sound features to the server 160, as will be discussed below.

[0081] As described above, the detection device 650 can actively monitor the hemodialysis treatment. When the pump of the hemodialysis machine 602 fails, it may generate an operating sound. The detection device 650 processes the detected operating sound into a sound feature (e.g., one or more discrete data elements). Then, the processor of the detection device compares the detected sound feature with a predetermined sound feature provided by the server 160. In this case, the server 160 provides a predetermined sound feature representing the operating sound of the pump.

[0082] When the detection device 650 detects a sound feature that matches a predetermined sound feature provided by the server 160, the detection device 650 can deploy intervention. In the example of the pump running sound, the detection device 650 can automatically detect whether the system is in pre-treatment (e.g., a setting for treatment), treatment or after treatment. If the system is in pre-treatment or after treatment, the detection device 650 can cause the hemodialysis machine to restart the system or run a diagnostic program. If the system is in treatment, the detection device 650 can cause the hemodialysis machine to run a diagnostic program after the treatment is completed. Because the detection device 650 can communicate directly with the hemodialysis machine, the detection device no longer requires the patient to restart the system or manually run a diagnostic program.

[0083] In some embodiments, the processor of the dialysis machine deploys intervention. For example, when the detection device 650 detects a sound signature that matches a predetermined sound signature provided by the user, the processor of the dialysis machine can restart the system or run a diagnostic program.

[0084] Figure 7A flow chart of a method for detecting sound and deploying intervention is shown. The method starts with detecting sound. For example, a detection device may detect sound as described above (702). The sound is processed to determine a sound feature (704). For example, the sound may be processed into one or more discrete data elements representing the sound feature. Next, the detected sound feature determines that the detected sound feature matches a predetermined sound feature (706). For example, the detected sound feature is compared with a list of predetermined sound features to determine that the detected sound feature matches a predetermined sound feature provided by a database or server. Determine appropriate intervention based on the predetermined sound feature (708). For example, each predetermined sound feature may be linked to an intervention (e.g., by a lookup table). The method may include determining a treatment stage (e.g., pre-treatment, during treatment, after treatment) (710) after determining that the detected sound feature matches the predetermined sound feature. For example, the method may include initiating an audio prompt to a patient, a caregiver, or a clinician, and requesting more information from the patient or clinician to identify whether the system is in pre-treatment (e.g., set for treatment), during treatment, or after treatment. After determining the treatment stage, deploy intervention (712). For example, the deployment intervention may include sending an audio prompt to the patient, caregiver, or clinician to restart the system or run a diagnostic program. In another example, the deployment intervention may include restarting the system or running a diagnostic program.

[0085] The server described herein can be updated to include more sound features and interventions. For example, because the detection device actively listens during treatment, the sound features that occur during the treatment process can be uploaded to the database. Before the sound features are uploaded to the server, the sound can be processed on the detection device (e.g., echo cancellation, noise reduction, digital processing to suppress unwanted features) to anonymize the sound features. If no errors or alarms occur during the treatment process, the sound features of the treatment can be uploaded to the database and stored as "normal operation" sound features. For example, the sound features of a normally operating machine can be uploaded and stored as normal operation sound features. If it is determined during or after treatment that the machine has failed, the sound features that occurred during the treatment process can be uploaded to the database and used to determine new predetermined sound features indicating a failure.

[0086] To identify new predetermined sound signatures, the server may process (e.g., via artificial intelligence or machine learning as described below) the sound signatures of the treatment. The server may compare the sound signatures of the malfunctioning treatment with the sound signatures of stored "normal operation" sound signatures. If a certain sound signature often precedes a certain type of failure, then the sound signature may indicate that type of failure. For example, if a blood pump makes an operating noise before a failure, then the sound signature of the operating noise may indicate a blood pump failure. Once the server identifies a sound signature that indicates a certain type of failure, the server may update the list of predetermined sound signatures to include the new sound signature.

[0087] One or more machine learning techniques may be employed. For example, supervised learning techniques may be implemented, wherein training is based on expected outputs known for inputs. Supervised learning may be considered an attempt to map inputs to outputs and then estimate outputs for previously unseen inputs (newly introduced inputs). Unsupervised learning techniques may also be employed, wherein training is provided based on known inputs but unknown outputs. Reinforcement learning techniques may also be used, wherein the system may be considered to learn from the results of actions taken (e.g., input values ​​are known and feedback provides performance measurements). In some arrangements, the techniques implemented may employ two or more of these approaches.

[0088] Figure 8 A flow chart of a method for determining a new sound signature indicating a malfunction is shown. First, sounds detected during treatment are processed to anonymize the sounds and convert the sounds into sound signatures (802). For example, as described above, a detection device may process sounds detected during treatment. Then, sound signatures are received (804). The server determines that a malfunction or failure occurred during the treatment process (806). For example, the server may receive confirmation that a malfunction occurred during treatment, such as at a particular time. The server then analyzes the sound signature, such as when the malfunction occurred or before the malfunction occurred, to determine a sound signature that occurred before the malfunction (808). For example, the server may compare the sound signature of the malfunctioning treatment with the sound signature of a stored "normal operation" sound signature. If the sound signature indicates a certain type of malfunction, the sound signature may be stored in a list as a predetermined sound signature (810).

[0089] Although the detection device has been described as actively listening for the duration of the treatment, in some embodiments, the detection device actively listens only for portions of the treatment. For example, the detection device may actively listen for a particular phase of the treatment (e.g., pre-treatment, during treatment, post-treatment). In some embodiments, the detection device samples the sound during the entire treatment (e.g., the detection device actively listens for a few seconds, such as every 15 seconds).

[0090] Although specific sounds have been described, other sounds that occur during treatment can be used as predetermined sound features. For example, the rattle or shake of the machine can be stored as a predetermined sound feature. In another example, an alarm or notification generated by the machine can be stored as a predetermined sound feature. In another example, the irregular sound of an electromechanical component (e.g., a fan, a servo motor, a heater, etc.) can be stored as a predetermined sound feature. In another example, the power sequence of the machine (e.g., power on or off) can be stored as a predetermined sound feature. The whooshing sound of blood, liquid flow, or air flow can also indicate a fluid line leak and can be stored as a predetermined sound feature. For example, dialysate, blood, or replacement fluid may leak from the liquid line of a malfunctioning dialysis machine.

[0091] While specific interventions have been described, other interventions may also be utilized. One example of an intervention is an audio prompt to a local patient, caregiver, or clinician, including instructing the patient to attempt to triage the problem or call technical support or their nurse from their clinic. Another example of an intervention is an alert or notification to a clinical network dashboard sent to healthcare provider staff. Yet another example of an intervention is an alert or notification sent to medical device manufacturer staff (e.g., technical or customer support staff). Another example of an intervention is a prompt for incoming communications (e.g., voice call, video chat, text message) so that the patient can communicate with medical device manufacturer staff or clinicians. Another example of an intervention is a remote software update to a treatment system. Another example of an intervention is an application notification sent to the patient. For example, a patient may have an application on a smartphone that allows the patient to receive push notifications on the home screen of the smartphone (e.g., Fresenius TM Fitness tracker apps).

[0092] Although certain types of detection devices have been described, other types of detection devices may also be used. Some examples of detection devices are smartphones, smart watches, tablets, laptops, smart home devices, Fresenius Kinexus TM A gateway, a smart TV, etc. A device including a microphone and a processor and capable of communicating with a server may be used as the detection device as described above.

[0093] Although the predetermined sound signature has been described as being identified by a server (e.g., through machine learning or artificial intelligence), in some embodiments, the predetermined sound signature may be identified and uploaded by a nurse, clinician, medical device manufacturer, or patient. For example, the known event and sound signature may be verified by customer service, technical support, clinicians, marketing experts, or R&D personnel.

[0094] In certain embodiments, the speaker of the hemodialysis machine can send an audio prompt to the user. For example, the detection device can send a signal to the hemodialysis machine so that the speaker of the hemodialysis machine sends an audio prompt. For example, the hemodialysis machine can use a speaker to send an audio prompt to the patient, a caregiver, or a clinician to request more information from the patient or clinician to identify whether the system is in pre-treatment (e.g., set for treatment), treatment, or after treatment. The hemodialysis machine can send an audio prompt to the patient, a caregiver, or a clinician to restart the system or run a diagnostic program. The hemodialysis machine can also send an audio prompt to the patient, a caregiver, or a clinician to run a diagnostic program after the treatment is completed.

[0095] In some embodiments, the patient can manually upload the sound signature using the detection device. For example, the patient can use a smartphone application (e.g., the Fresenius "My Companion" patient application) to record the sound signature and submit it to the server. The server can then process the sound signature as described above. This may be useful if the patient identifies irregular noises during treatment.

[0096] In some embodiments, the cloud-based voice arrangement platform can be used to distribute or publish sound features or communicate between the detection device and the server. In addition, the detection device can communicate with the server through one or more hard-wired connections (e.g., serial cables, parallel cables, etc.). Wireless technologies such as infrared (IR), radio frequency (RF), Bluetooth, wireless Ethernet or other electromagnetic link technologies can also be used. One or more protocols, transmission standards and data formats can also be used to transfer information between the detection device and the server.

[0097] In some embodiments, storing the predetermined sound signature may include storing additional data about the sound signature, including, but not limited to, a timestamp of occurrence, identification of the medical treatment device, and / or location of occurrence.

[0098] Although the above detection devices and servers have been described as part of a PD system and a hemodialysis system, these types of devices can be used in any of a variety of other types of medical treatment systems and medical fluid pumping systems. Some examples of medical treatment systems include life measurement devices, fluid heaters, and other systems with fluid lines. Other examples of medical fluid pumping systems in which the devices and servers described herein can be used include hemofiltration systems, hemodiafiltration systems, single collection systems, and infusion pump systems.

Claims

1. A method for monitoring a medical treatment device, the method comprising: detecting the sound of the medical treatment device using a detection device; determining whether the sound is an irregular sound; as well as An intervention is deployed upon determining that the sound is an irregular sound.

2. The method according to claim 1, wherein: Deployment interventions include displaying notifications.

3. The method according to claim 2, wherein: The notification is a push notification on a smartphone.

4. The method according to any one of claims 1 to 3, wherein: Deploying the intervention includes initiating an audio prompt on a device associated with a patient of the medical treatment device.

5. The method according to any one of claims 1 to 4, wherein: Deploying the intervention includes initiating a telephone call to a telephone associated with a patient of the medical treatment device.

6. The method according to any one of claims 1 to 5, wherein: Deploying the intervention includes initiating an audio prompt on a device associated with a patient and / or clinician of the medical treatment device to restart the medical treatment device.

7. The method according to claim 6, wherein: Deploying an intervention includes initiating the audio prompt on the device associated with the patient and / or clinician to restart the medical treatment device after confirming that the medical treatment device is in a pre-treatment stage.

8. The method according to claim 7, wherein: The irregular sound is associated with a pump of the medical treatment device.

9. The method according to any one of claims 1 to 8, wherein: Determining that the sound is an irregular sound includes comparing the sound with sound samples stored in a database.

10. The method according to claim 9, wherein: The sound samples are uploaded to the database and recognized through machine learning.

11. The method according to any one of claims 1 to 10, wherein: The irregular sound indicates a malfunction of the medical treatment device.

12. The method according to claim 11, wherein: The irregular sound indicates a disconnected vein.

13. The method according to claim 11, wherein: The irregular sound indicates a pump malfunction.

14. The method according to any one of claims 1 to 13, wherein: The medical treatment device is a hemodialysis machine.

15. The method according to any one of claims 1 to 12, wherein: The medical treatment device is a peritoneal dialysis machine.

16. A medical treatment system comprising: Medical treatment devices; a detection device configured to detect sounds of the medical treatment device; A processor is configured to determine whether the sound is an irregular sound and deploy an intervention when the sound is determined to be an irregular sound.

17. The medical treatment system of claim 16, wherein: The intervention includes initiating an audio prompt on a device associated with a patient of the medical treatment device to restart the medical treatment device.

18. A medical treatment system according to claim 16 or claim 17, wherein: The detection device comprises a microphone.

19. The medical treatment system according to any one of claims 16 to 18, wherein: The processor is configured to determine whether the sound is an irregular sound by comparing the sound with a sound database.

20. The medical treatment system according to any one of claims 16 to 19, wherein: The sound indicates a malfunction of the medical treatment device.

21. The medical treatment system of claim 20, wherein: The sound indicates a disconnection of the vein.

22. The medical treatment system of claim 20, wherein: The sound indicates a pump malfunction.

23. The medical treatment system of claim 20, wherein: The sound is the rattling sound of the housing of the medical treatment device.

Images