System and method for monitoring an infusion pump

Abstract

System (2) for monitoring an infusion pump (6), comprising: an infusion pump (6) having at least one sensor (10, 14, 18) for detecting sensor data; a monitoring server (26) having a monitoring module (26) with a receiving unit (36); a transmission unit (30) for transmitting the sensor data to a receiving unit (36) of the monitoring module (26); wherein the monitoring module (26) is configured to determine a state of the infusion pump (6) on the basis of the sensor data and to derive therefrom a safety-related classification of the infusion pump (6).

Description

Technical Field

[0001] The present invention relates to a system and method for monitoring an infusion pump. Background Art

[0002] An infusion pump is a medical device that enables targeted and automatic administration of fluids, such as medications or nutrients. The safety and specific mode of operation of these pumps are crucial, as a malfunction can have serious consequences for the patient.

[0003] To ensure the safety of infusion pumps, various safety-related tests are currently performed. The tests are intended to ensure that the pumps function properly during production and throughout the life of the device. Known technical safety checks (TSCs) focus on identifying electrical disturbances and mechanical defects of infusion pumps and include safety-related work.

[0004] The technical safety check is performed by a maintenance technician specially trained for this purpose to maintain the device and keep it functional. The maintenance process is complex and can be costly. During the test, the infusion pump is placed in maintenance mode, making it unavailable for use in regular hospital operations.

[0005] A system is known from WO 2021 / 243068 A1, which enables monitoring of components via sensors in conjunction with a server and, in the event of a fault, sends a message to the relevant device and displays maintenance instructions on the device.

[0006] A disadvantage of known technical safety tests is that they require the infusion pump to be taken out of service and require the use of special technicians, or they support maintenance or repair of the infusion pump only after a fault has been detected. Summary of the invention

[0007] Therefore, the object of the present invention is to propose a system and method for simplifying and automating technical safety checks. In addition, a computer product and a training data set should also be proposed.

[0008] With regard to the system, the object is achieved according to the invention by means of a system for monitoring an infusion pump having the features of claim 1. The system comprises an infusion pump having at least one sensor for detecting sensor data, a monitoring server having a monitoring module with a receiving unit, and a transmission unit for transmitting the sensor data to the receiving unit of the monitoring module, wherein the monitoring module is configured to determine a state of the infusion pump based on the sensor data and to derive therefrom a safety-related classification of the infusion pump.

[0009] Advantageous embodiments of the invention are the subject matter of the dependent claims.

[0010] The invention is based on the consideration that an infusion pump should be available in medical operation as uninterruptedly as possible. Due to necessary technical safety monitoring or safety checks by maintenance technicians, the pump needs to be shut down, during which time it is unavailable, so that a replacement device must be organized and organizational measures must be taken.

[0011] As has already been recognized, a technical safety check by a service technician can be avoided by means of an automated technical safety check in that the state of the infusion pump is evaluated based on sensor data. This is particularly possible because sensors already present in the infusion pump can be used for this purpose. If the data meet preset or learned conditions, in particular when selected or all sensor data are within preset tolerances, the safety check can be considered to be satisfied, so that a manual check can be omitted and the infusion pump can remain in operation.

[0012] Here, the sensor data are integrated to create a safety-related classification. With the aid of the sensor data, a prediction of the degradation of the infusion pump can be created. Then, preferably in a further step, the sensor data are integrated to create a safety-related classification. The safety-related classification preferably includes a decision as to whether the device is still considered safe and whether the safety check has been passed.

[0013] Therefore, the safety-relevant classification is preferably a different or further classification of the data. For example, the entire system or infusion pump can still be considered safe, although the "prediction score" has been proposed to be maintained. Another situation can also be considered: that is, the "prediction score" has not yet been achieved, but the entire system or infusion pump has been considered unsafe in the sense of the safety-relevant classification.

[0014] Thus, the system can not only perform predictive maintenance ("predictive maintenance") on infusion pumps, but can also perform safety-related classification within the scope of safety-related checks (STK) and help the user to automate said classification without any user intervention.

[0015] Advantageously, the safety-related classification comprises a prediction of malfunction and / or proper operation of the infusion pump during operation time.

[0016] The safety-related classification is preferably performed in cycles, wherein the safety-related monitoring of the infusion pump is considered to be passed if the infusion pump has functioned normally during the last cycle. In this case, the safety-related manual test by the service technician can be omitted, so that the infusion pump continues to be available for its operating state.

[0017] The transmitting unit preferably transmits the detected sensor data to the receiving unit substantially continuously or at regular time intervals. Depending on the data gaps, interpolation (linear or spline) is preferably performed and the missing data values ​​are approximated, while in the case of larger data gaps, the trend analysis is advantageously aborted and a reinitialization is performed. Conversely, this means that the previously used data are declared invalid and the trend analysis cannot be performed.

[0018] In a preferred embodiment, the monitoring module comprises an adaptable and / or learnable model for safety-related classifications. In this way, the model can be adapted to the specific technical situation, i.e. a specific pump and a specific sensor, and the criteria for a proper functioning of the infusion pump can be adapted over time to any newly acquired knowledge.

[0019] The algorithm of the model is advantageously a prediction scheme, in particular comprising at least one linear filter model and / or a statistical model, in particular comprising at least one neural network and / or a Gaussian mixture model.

[0020] In particular, the model can include different models for different groups of sensor data. In a preferred embodiment, a plurality of models are therefore provided, each of which is dedicated to a selection of the entire sensor data. Thus, certain sensor data can be binary, wherein the sensor data value directly leads to a classification of a fault or failure. Other sensor data or combinations thereof can lead to more complex classifications with intermediate levels.

[0021] For each sensor there is advantageously a different model, which is initially uncorrelated. Correlation is only made in the following way: in the event of a malfunction (i.e. defect), the infusion pump is typically marked as defective, or in the case of a "recoverable" error (e.g. display calibration), a calibration is started / initiated by the user.

[0022] The monitoring module is preferably configured to set the parameters of the model in the adaptation and / or learning phase. That is to say, the parameters of the model are confirmed with the aid of a training data set. The training can be performed cyclically. In the case of a neural network, the network is advantageously trained by supervised learning with the aid of a training data set. In a preferred embodiment, the learning success of the neural network is then checked with the aid of a test data set. If a preset test criterion is not met, the neural network can be trained with the aid of changed training parameters (e.g., learning rate) and / or a changed training data set. This can also be applied analogously to a Gaussian mixture model. Here, the learning success is also checked with the aid of other data.

[0023] In the case of a syringe pump, the monitoring module preferably detects and analyses (ie evaluates with the aid of a model) sensor data from the following group:

[0024] The jaw angle, which is related to the syringe used and its length / its diameter as a function of time;

[0025] Syringe length / syringe diameter of the syringe used as a function of time;

[0026] • Pressure values ​​during insertion of the syringe and / or during infusion.

[0027] In the case of a volumetric pump / peristaltic pump, the monitoring module preferably detects and analyzes (ie evaluates with the aid of a model) sensor data from the following group:

[0028] Moving the door position sensor by checking the resistance value over time and correlating that value to the motor steps;

[0029] color detection of the inserted hose, in particular based on the clamp of the hose, wherein the measured color value is evaluated by means of confirmation / change by the user;

[0030] · Detect pressure sensor values ​​over time to check if there is a drift / change over lifetime.

[0031] In the case of a peristaltic pump / volumetric pump with display or a syringe pump with display, the monitoring module preferably detects and analyses (ie evaluates with the aid of a model) sensor data from the following group:

[0032] The contact sensor value as a function of time at a defined position on the display, for example on a "start infusion" switch face or in an input field;

[0033] • Exchange battery information such as state of charge as a function of time, battery capacity, number of charge cycles through device models.

[0034] The system preferably comprises transmitting means for transferring information from the monitoring module to the infusion pump. Advantageously, the receiving unit and the transmitting means are integrated or combined, ie the same unit receives and sends.

[0035] The monitoring module is preferably designed to monitor whether the sensor data of the at least one sensor are within a preset and / or learned nominal range, wherein the monitoring module is designed to send instructions to the infusion pump by means of the transmission device and / or to initiate a maintenance procedure of the infusion pump, preferably a display calibration. This is preferably done in the case of sensor data regarding properties of the infusion pump, which can be changed, in particular temporarily, by a medical specialist during a maintenance step. For example, a touch screen or a touch display can be recalibrated.

[0036] The instructions and / or steps of the maintenance procedure are preferably displayed on a display of the infusion pump. Additionally or alternatively, sound signals or voice messages may also be used.

[0037] The transmitting unit and the receiving unit are preferably designed for wired or wireless communication, in particular by means of radio, mobile radio, W-Lan, Bluetooth. In this way, no additional cables or lines need to be laid, which makes the working area of ​​the infusion pump clearer.

[0038] The system advantageously also comprises a health data server configured to receive data from the monitoring module and / or to transmit data to the monitoring module. The server preferably receives data continuously, i.e. whenever a change occurs. The pressure sensor value is continuously transmitted and monitored, for example while treatment / delivery is ongoing.

[0039] However, when inserting a syringe into a syringe pump, the claw angle and its diameter are measured primarily during the insertion process. Then, measurements are taken and reported at irregular intervals (e.g., at intervals of 50 milliseconds to 5 seconds), but always when a new syringe is inserted. The same applies to the air sensor of a peristaltic pump. Here, the data are measured continuously in the case of a disposable item being inserted. If possible, the data are delivered in packages to a server. On the one hand, this occurs event-controlled, e.g., when the flow rate changes, a syringe / tube is inserted or replaced, and in any case preferably every 5 seconds.

[0040] The health data server is advantageously designed as a cloud server.

[0041] The monitoring server advantageously comprises a trend analysis module, which is configured to perform a trend analysis based on the transmitted sensor data. The trend analysis module is preferably configured in the monitoring server in the form of software and / or hardware.

[0042] Based on the trend analysis, the trend analysis module preferably performs a life assessment of the infusion pump. Errors can be identified early and reacted to by the trend analysis. Preferably, the technician is informed of preventive maintenance and, if necessary, a message is sent to the pump to send it for repair. However, the latter depends on whether the pump is still running at the patient or it is not in use. In the latter case, it can be considered to provide the pump with a maintenance reminder. In the case of a touch display, in a preferred embodiment, it can be considered to send a "please recalibrate" field so that the caregiver can proceed accordingly. This is only done outside of an ongoing treatment. Here, in particular by progressive prediction, it can be recognized early that the error will worsen over time or the expected operating time / lifetime of the infusion pump can be adapted.

[0043] As an example, a trend can be identified that the touch display needs to be recalibrated. This can be identified, for example, as a trend that the pixel positions corresponding to the user's touch gestures for a switch face whose position is known are constantly changing. For example, the user must always press the display further upwards in order to activate the switch face.

[0044] Regarding the method, according to the present invention, the above object is achieved by a method having the following steps: detecting sensor data of an infusion pump; inputting the sensor data into a mathematical model; and classifying the state of the infusion pump by means of the mathematical model. The method is implemented by a computer.

[0045] In a preferred embodiment, the method comprises an adaptation and / or training step for the mathematical model, in which the model is adapted and / or trained with the aid of a training data set.

[0046] The training method of the statistical model preferably only includes sensor data that have caused a fault or are in an error-free state. It is known a priori whether it is a good situation (no error) or a bad situation (fault). The mathematical / statistical model is trained accordingly for this. An adaptation phase can also be provided, in which the statistical model is adapted and improved by further data (measured values) from the field. This is advantageous when the sensor data has a high variance. The variance can be incorporated into the learned model by retraining (i.e. adaptation), thereby producing greater robustness. It is also possible to adapt to reality, i.e. to continuously improve / adjust the model with the help of real data.

[0047] The training data set advantageously comprises data of at least one sensor and an associated classification. Advantageously, more than one sensor is included in the consideration. The sensor values ​​are not correlated and must be considered individually. Good cases and bad cases are needed in the training data in order to achieve the classification.

[0048] The mathematical model advantageously comprises at least one model from the following group: a linear filter model and / or a statistical model, in particular comprising at least one neural network and / or a Gaussian mixture model.

[0049] The method preferably comprises the method step of monitoring sensor data of at least one sensor, wherein as soon as the sensor data leave a preset and / or learned nominal range, a maintenance instruction is displayed on a display of the infusion pump.

[0050] Here, maintenance instructions are basically instructions for recalibrating the display or touch calibration. In the case of other defects, the corresponding maintenance instructions can be a reminder that the pump should be handed over to a maintenance technician (because there is a defect).

[0051] As regards computer products, according to the invention, the above objects are achieved by a computer program product comprising instructions which, when executed on a computer, perform the above method.

[0052] With regard to the training data set for training the neural network in the method with the above training step, the above object is achieved according to the invention with the aid of a training data set having the following structure. The training data set preferably contains only sensor data that have caused a fault or belong to an error-free state. It is known a priori whether it is a good situation (no fault) or a bad situation (fault). The mathematical / statistical model is trained accordingly for this.

[0053] In the adaptation phase, further data (measurements) from the field can be used to adapt and improve the statistical model. This is particularly advantageous when the sensor data has a higher variance. The variance can be incorporated into the learning model by retraining (i.e. adaptation), and thereby greater robustness is generated. The training data set includes to a certain extent prior knowledge about defective and error-free sensor data. In addition, sensor data from different numbers of infusion pumps (hose pumps or syringe pumps) are preferably used for training in order to cover sensor divergence due to component tolerances. In addition, sensor data from infusion pumps of different batches and different ages are advantageously used in order to better / more accurately describe the divergence of the sensor data.

[0054] The present invention also relates to a system for monitoring a flow pump or a dialysis machine, comprising a flow pump or a dialysis machine having at least one sensor for detecting sensor data, a monitoring server having a monitoring module with a receiving unit, and a transmission unit for transmitting the sensor data to the receiving unit of the monitoring module, wherein the monitoring module is configured to determine the state of the flow pump / dialysis machine based on the sensor data and to derive therefrom a safety-related classification of the flow pump / dialysis machine. The preferred embodiments described above for the system for monitoring an infusion pump in the claims or the specification are also preferred embodiments of the system for monitoring a flow pump or a dialysis machine, as long as they are technically directly or equivalently transferable to a person skilled in the art.

[0055] The advantages of the invention are in particular that an automated system is provided in order to detect faults at an early stage and / or to prolong the maintenance of the device based on a trend analysis supported by sensor values. In this way, the infusion pump can remain in operation for as long as possible. By monitoring the infusion pump and / or the trend analysis performed, maintenance intervals can be postponed or brought forward, thereby ensuring the functional use of the infusion pump. Since maintenance must usually be carried out every two years in Germany, continuous monitoring makes it possible to carry out checks based on the data. This saves time and correspondingly money. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Embodiments of the invention are explained in more detail with reference to the accompanying drawings. In the highly schematic view:

[0057] Figure 1 A preferred embodiment of a system is shown, wherein the system has an infusion pump, a monitoring server, and a health data server;

[0058] Figure 2 Another preferred embodiment of the system is shown, the system having an infusion pump, a monitoring server and a health data server;

[0059] Figure 3 showing a display of the infusion pump in a first state;

[0060] Figure 4 Shows the second state according to Figure 3 Display;

[0061] Figure 5 Shows the third state according to Figure 3 Display; and

[0062] Figure 6 A flow chart showing a preferred embodiment of the method.

[0063] In all the figures, identical parts are provided with identical reference numerals. DETAILED DESCRIPTION

[0064] Figure 1 The system 2 shown in the figure comprises an infusion pump 6 having a plurality of sensors 10, 14, 18. The system 2 comprises a monitoring server 22 having a monitoring module 26, which is implemented in the monitoring server 22 in a hardware and / or software manner. The system 2 further comprises a transmission unit 30, which is connected to the sensors 10, 14, 18 or a processor, and which is configured or designed to transmit data detected by the sensors or sensor data to a receiving unit 36 ​​of the monitoring module 26, wherein the sensors are connected to the processor.

[0065] The system 2 further comprises a health data server 38 , which is preferably designed in the present case as a cloud server 40 , wherein the monitoring server 22 and the cloud server 40 are designed for data transmission from the monitoring server 22 to the cloud server and / or from the cloud server 40 to the monitoring server 22 .

[0066] exist Figure 2Another preferred embodiment of the system 2 is shown in . The system 2 comprises an infusion workstation 46, which comprises an infusion pump 6 as well as a rack system 50 and a communication system 54. In the embodiment, the infusion workstation 46 has a transmission unit 30. The communication system 54 may have a transmission unit 58 for exchanging data with the monitoring server 22. The infusion workstation 46 is configured to accommodate a plurality of infusion pumps 6, in particular up to 24 infusion pumps, in the rack system 50, which forms a accommodating system for the infusion pumps 6. Therefore, in the configuration, a common transmission unit 30 is provided for all infusion pumps 6 arranged in the rack system 50. Each pump preferably sends its data to the server 22 individually via its communication unit. Each pump has a unique identification, whereby it can be associated accordingly.

[0067] Figure 1 and 2 The two systems 2 shown in FIG. 2 each have a transmission device 62 for transmitting information from the monitoring module 26 to the infusion pump 6. In the present case, the receiving device 36 and the transmission device 62 are integrally formed. The monitoring servers 22 of the systems 2 also each include a trend analysis module 28.

[0068] The sensors 10, 14, 18 may be sensors from the following group:

[0069] -Sensor for measuring claw angle

[0070] -Sensor for measuring syringe diameter

[0071] -Pressure sensor for measuring the pressure inside the syringe

[0072] - Door sensor for volumetric or peristaltic pumps

[0073] -Color detection sensor for peristaltic pumps

[0074] -Touch display sensor

[0075] -Sensor for battery status

[0076] - Clamp on the equipment side of the peristaltic pump (to monitor the hose for possible pinching, e.g. when the door is open)

[0077] -Sensor for checking the hose clamp inserted in the peristaltic pump

[0078] - Connection to power supply (yes / no)

[0079] - Location of the piston flow meter in the syringe pump.

[0080] The nurse call is preferably monitored, ie whether the signal is actually applied at the connector. The LEDs (red, yellow, white) are preferably monitored. For this purpose, the voltage values ​​are checked for plausibility / drift and fed to the model accordingly.

[0081] The monitoring module 26 Figure 1 and 2 In the embodiment shown in , the monitoring module 26 is configured to perform a state classification of the infusion pump 6 based on the sensor data input by the sensors 10, 14, 18. The monitoring module 26 is configured to be adapted or trained in order to perform the proposed classification of the state. In an embodiment in which the model is a statistical model, in particular a neural network, a training data set containing the sensor data is provided. In addition, a test data set is provided to test the quality or accuracy of the trained model.

[0082] Here, supervised learning is preferably performed to ensure that the model is trained correctly. Reinforcement learning can also be used, provided that the underlying Markov chain is understandable / comprehensible, i.e., provided that the Markov model is validated in terms of conditions. An exemplary training data set contains measurements for identifying syringe sizes. When the syringe is inserted into the syringe pump, the diameter is measured based on the resistance value. The resistance value varies on the one hand according to the syringe size (e.g., the diameter of a 50 ml syringe is larger than the diameter of a 20 ml syringe), and on the other hand according to the insert. It is also known: what is the nominal diameter of the syringe and whether it is a good mode or a bad mode.

[0083] A similar example can be determined for the position of the door attitude. The exemplary system has a total of 3 different positions of the door attitude (open, closed and "parking position"). Due to component tolerances, the measured values ​​also vary around the nominal values ​​of the positions. Here, good and bad patterns can also be generated for training by threshold values.

[0084] In addition, the touch position on the display or screen can be selected as an input variable relatively easily. The training data set learns the position of each graphical element on the screen and the variance of the data point from the nominal value. If the measured value is within the range, the measured value is considered good. If the measured value is outside the range, it is marked as an error.

[0085] The main task of the current technical safety check is to check whether the button reacts to the touch or whether the alarm indication of the infusion pump (such as yellow status light, red status light, green infusion status light) is working properly. Currently, the equipment must undergo a safety check every two years. The present invention makes it possible to outsource the individual checks to professionals (such as checking LEDs or testing touch screens) in order to save money and time. The second aspect of the present invention is that due to the continuous monitoring of the device and its measured values, it can be very clearly identified whether an error is quietly occurring or not at all. There is the possibility of automating the part and thus avoiding the need for a dedicated safety check, because the data is checked continuously (or at shorter intervals than the safety check setting).

[0086] Therefore, one aspect of the present invention is to send information messages to the infusion pump 6 through the monitoring server 22, and thereby notify the user or medical expert. Figure 3 As shown in , the medical staff is informed by a message 70 on the display 66 of the infusion pump 6, which is designed as a touch display: from the next day, a message for checking the status light is displayed on the pump that is not infusing. The notification is only displayed when the infusion pump is inactive.

[0087] Regardless of whether such a message is posted, the monitoring server 22 displays a message 70 on the display 66 of the infusion pump 6 that the user should check the status light, as in Figure 4 , a red warning light 74 is shown by way of example.

[0088] The status lights are activated in sequence and the user must confirm that the corresponding status light is on. Confirmation is performed on the touch screen in order to also check the reaction of the display 66 to the touch input. For this purpose, two switch surfaces 78, 82 are faded in on the display, one switch surface 78 representing the answer "yes" and the other switch surface representing the answer "no".

[0089] After the corresponding status light has been displayed and confirmed, the user receives a short message to press a hard key button (i.e. not a button on the touch display, but a mechanical button) to exit the dialog. By confirming the command using the touch screen and the hard key button, the mandatory functions of the user interface are also checked. The described dialog, which is carried out by the user via the displayed commands and button confirmations, can also be supplemented by an acoustic signal for checking the loudspeaker and / or the backup alarm.

[0090] Finally, after all tests have been completed, the technical safety check is completed and a corresponding message 70 is displayed to the user on the display 66 , which the user can confirm by pressing a stop button 86 , which is currently designed as a hard key button, thereby completing the safety check.

[0091] exist Figure 6A preferred embodiment of the method for monitoring an infusion pump 6 is shown in FIG. In a first method step 90, sensor data of the infusion pump 6 are detected, which are measured by sensors. The sensor data are detected continuously or at regular intervals. In a second step 94, which is executed at least partially in parallel with the first step 90, the sensor data are input into a mathematical model, which is performed automatically and at regular intervals. The model is a linear filter model and / or a statistical model, in particular comprising at least one neural network and / or a Gaussian mixture model.

[0092] In a third method step 98, which is executed at least partially in parallel with method step 94, the state of the infusion pump 6, in particular the safety-related state, is classified. The specific classification can be designed differently depending on the model. Thus, in the case of a simple linear model, the classification can be a binary decision: the state is acceptable / the state is unacceptable from the safety point of view; the infusion pump must be maintained or repaired.

[0093] The sensor data is measured and compared with the model. This means: based on the sensor values ​​and previously measured values, an assessment / prediction of the "device defective" or "device ready" state is obtained (via filtering methods or machine learning).

[0094] If the device is marked as "device ready", nothing happens at all. The device can continue to be used. However, if a "device defective" status is estimated, this will be displayed on the device in the form of a message (in the form of an error notification). In addition, a message to the service technician is created on the server. An email or other message can also be sent to the service technician via the server for the purpose of repairing the infusion pump.

[0095] Reference numerals list

[0096] 2 System

[0097] 6 Infusion Pump

[0098] 10 Sensors

[0099] 14 Sensors

[0100] 18 Sensors

[0101] 22 Monitoring Server

[0102] 26 Monitoring Module

[0103] 28 Trend Analysis Module

[0104] 30 transfer units

[0105] 36 Receiving Unit

[0106] 38 Health Data Server

[0107] 40 Cloud Servers

[0108] 46 Infusion Workstation

[0109] 50 Rack System

[0110] 54 Communication System

[0111] 58 transfer units

[0112] 62 Transmission Device

[0113] 66 Display

[0114] 70 messages

[0115] 74 Red warning light

[0116] 78 switch surface

[0117] 82 switch surface

[0118] 86 Stop button

[0119] 90 steps

[0120] 94 Steps

[0121] 98 Steps

Claims

1. A system (2) for monitoring an infusion pump (6), comprising: An infusion pump (6) having at least one sensor (10, 14, 18) for detecting sensor data; A monitoring server (26) having a monitoring module (26) with a receiving unit (36); a transmitting unit (30) for transmitting the sensor data to a receiving unit (36) of the monitoring module (26); The monitoring module (26) is configured to determine a state of the infusion pump (6) based on the sensor data and to derive therefrom a safety-related classification of the infusion pump (6).

2. The system (2) according to claim 1, wherein the safety-related classification includes a prediction of a failure and / or a normal operation of the infusion pump (6) during operating time.

3. The system (2) according to claim 2, wherein the safety-related classification is performed cyclically, and wherein: If the infusion pump (6) operates normally in the last cycle, it is considered that the safety-related monitoring of the infusion pump (6) has passed.

4. The system (2) according to any one of claims 1 to 3, wherein the transmitting unit (30) transmits the detected sensor data to the receiving unit (36) substantially continuously or at regular time intervals and / or The monitoring module (26) comprises an adaptable and / or learnable model for safety-related classifications.

5. System (2) according to claim 4, wherein the algorithm of the model is a prediction scheme, in particular comprising at least one linear filtering model and / or a statistical model, in particular comprising at least one neural network and / or a Gaussian mixture model.

6. The system (2) according to any one of claims 1 to 5, wherein in the case of a syringe pump, the monitoring module (26) detects and analyzes sensor data from the following group: The jaw angle, which is related to the syringe used and its length / its diameter as a function of time; Syringe length / syringe diameter of the syringe used as a function of time; pressure values ​​during insertion of the syringe and / or during infusion, and / or, wherein in the case of a volumetric pump / peristaltic pump, the monitoring module (26) detects and analyses sensor data from the following group: Moving the door position sensor by checking the resistance value over time and correlating that value to the motor steps; color detection of the hose, in particular based on the clamp of the inserted hose, wherein the measured color value is evaluated by means of confirmation / change by the user; Detecting the pressure sensor value over time to check if there is a deviation / change over lifetime, and / or, wherein in the case of a peristaltic pump / volumetric pump with a display or a syringe pump with a display, the monitoring module (26) detects and analyses sensor data from the following group: The contact sensor value as a function of time at a defined position on the display, in particular the contact sensor value on the switch surface of "start infusion" or in the input field; Exchange battery information through device models, specifically the state of charge, battery capacity, and number of charge cycles as a function of time.

7. The system (2) according to any of the preceding claims, comprising a transmission device (62) for transmitting information from the monitoring module (26) to the infusion pump (6), and / or The monitoring module (26) is configured to monitor whether the sensor data of at least one sensor (10, 14, 18) is within a preset and / or learned nominal range, and to send instructions to the infusion pump (6) and / or initiate a maintenance procedure for the infusion pump (6) by means of the transmission device (62).

8. The system (2) according to claim 7, wherein the instructions and / or the steps of the maintenance procedure are displayed on a display of the infusion pump (6).

9. The system (2) according to any of the preceding claims, further comprising a health data server (38) configured to receive data from the monitoring module (26) and / or to transmit data to the monitoring module (26).

10. The system (2) according to any one of the preceding claims, wherein the monitoring server comprises a trend analysis module (28) configured to: perform a trend analysis based on the transmitted sensor data, and / or The trend analysis module (28) performs a life assessment of the infusion pump (6) based on the trend analysis.

11. A method for monitoring an infusion pump (6), comprising the following method steps: Detecting sensor data of the infusion pump (6); Inputting the sensor data into a mathematical model; Classifying the states of the infusion pump (6), in particular safety-related states, by means of the mathematical model. 12 . The method according to claim 11 , comprising an adaptation and / or training step for the mathematical model, in which the model is adapted and / or trained with the aid of a training data set.

13. The method according to claim 11 or 12, wherein the training data set comprises data of at least one sensor (10, 14, 18) and associated classifications.

14. The method according to any one of claims 11 to 13 comprises a method step of monitoring sensor data of at least one sensor (10, 14, 18), wherein maintenance instructions are displayed on a display (66) of the infusion pump (6) as soon as the sensor data leaves a preset and / or learned nominal range.

15. A computer program product comprising instructions which, when executed on a computer, perform the method according to any one of claims 11 to 14.

Images