Individualized patient therapy model using datamining and predictive algorithms

Abstract

A method for optimizing operation of an implanted and / or implantable medical device, IMD, comprises accessing medical data comprising operation parameters of a plurality of IMDs of a plurality of patients, determining at least one first operation parameter set for at least one IMD of at least one patient at least in part based on the medical data, and transmitting the at least one first operation parameter set to the at least one IMD.

Description

[0001] Applicant: BIOTRONIK SE & Co. KG

[0002] Our Reference: 20.136P-WO

[0003] Date: 08.08.2025

[0004] Individualized patient therapy model using datamining and predictive algorithms

[0005] The present disclosure relates to methods for optimizing operation of an implanted and / or implantable medical device (IMD), corresponding IMDs, data-processing devices, and computer programs.

[0006] Currently, adjusting IMDs to a patient’s needs is typically still strongly based on time-consuming and expensive face-to-face visits of the patient with a clinician. E.g., implanted pacemakers, cardiac monitors, neuro stimulators and similar IMDs may communicate with clinician user interface devices (“clinical programmers”) during such face-to-face visits. During the face-to- face visit the clinical programmer may typically provide a real-time view of physiological parameters and (technical) operation parameters. Further information on the status of care (e.g., expressed through physiological parameters of the patient) can be self-reported by the patient, typically during the face-to-face visit, or recorded by the IMD and / or other medical devices. In these sessions, the attending clinician typically cannot account for all available data, e.g., due to a lack of time and resources.

[0007] Such face-to-face visits are not only disadvantageous for the reasons above, they also cannot provide ideal care for the patient from a medical perspective: Many of these implanted and / or implantable devices can also transmit physiological and (technical) operation parameters periodically to a server (e.g., a server for remote patient monitoring). Clinicians who may, e.g., access this data through a secure website currently cannot exploit the full range of possibilities offered by the large amount of available data for optimized treatment of the patient.

[0008] Rather than providing proactive care, the current standard of care typically consists of the responsible clinician providing treatment once the patient becomes symptomatic, i.e., reactive care. There is limited capability in the state-of-the-art to automatically predict outcomes based upon quantitative data gathered from that patient, and / or via statistical correlation to the history of related patient populations. Similarly, there is limited capability in the state-of-the-art to automatically notify care providers of potential pending changes to the patient status. Further, little use is made of the medical data received via patient monitoring to improve the initial configuration of devices awaiting initial programming. Similarly, application of objective data to new product design is minimal.

[0009] As a result, there is still a need for improving methods for optimizing operation of an IMD, corresponding IMDs, data-processing devices, and computer programs.

[0010] The above need is at least partly met by the aspects as described herein.

[0011] A first aspect refers to a method for optimizing operation of an implanted and / or implantable medical device, IMD, the method comprising the steps of: accessing medical data comprising operation parameters of a plurality of IMDs of a plurality of patients; determining at least one first operation parameter set for the IMD of the patient at least in part based on the medical data; and transmitting the at least one first operation parameter set to the IMD.

[0012] Providing at least one first operation parameter set to the IMD based on the medical data (i.e., during operation of the IMD) may strongly improve the operation efficiency of the IMD, e.g., by operating it, as described herein, with an operation configuration according to the at least one first operation parameter set. The medical data may reflect current and / or local conditions of the IMD, the patient of the IMD, and / or the interplay between patient and IMD.

[0013] As understood herein, a set of elements may comprise a single element or, alternatively, it may comprise a plurality of elements. In detail, an operation parameter set may comprise one or more operation parameters, e.g., as described herein.

[0014] The method may, in some examples, further comprise determining at least one second operation parameter set. Therein, the determining of the first and second operation parameter set may, e.g., comprise: identifying at least a first and a second operation pattern at least in part based on the medical data; determining the first operation parameter set based on the first operation pattern and determining the second operation parameter set based on the second operation pattern; ranking the first operation pattern and the second operation pattern according to a first score associated with the first operation pattern and a second score associated with the second operation pattern; and defining a preference of the first operation parameter set and the second operation parameter set according to the ranking.

[0015] 20.136P-WO / 08.08.2025 Identifying different operation patterns based on the medical data allows determining which operation parameter set is to be preferred for given circumstances identified by the operation pattern. This way operation efficiency may be increased, as an operation parameter set can be selected based on said circumstances.

[0016] Identifying said operation patterns may be at least partially based on unsupervised learning and / or data mining, e.g., hierarchical clustering, k-means clustering etc. In addition, or alternatively, identifying said operation pattern may be at least partially based on self-supervised learning. In addition, or alternatively, identifying said operation patterns may be at least partially based on supervised learning. Generally, identifying said operation patterns may be at least partially based on decision trees and / or neural networks, etc.

[0017] Some examples for identifiable operation patterns may comprise one or more of the following (relating to an example wherein the IMD may be a neurostimulator): (1) There is a 100% chance of therapy failure when the lead impedance is above a first predetermined impedance threshold. (2) There is a 90% chance of therapy failure when there is a lead impedance changed by more than a second predetermined impedance threshold, e.g., . (3) There is a 100% chance of therapy failure when the battery usage exceeds a predetermined battery usage threshold. (4) There is a 75% chance of therapy failure when the patient-selected therapy usage rate is below a predetermined therapy usage threshold. (5) There is a 75% chance of therapy failure when the evoked action compound potential (ECAP) assessments fail to exceed activation threshold. (6) There is a 35% chance of therapy failure when the patient’s heart rate variability (HRV) is below a predetermined HRV variability threshold. (The specific values shown here are provided merely as examples.)

[0018] The method may, for example, further comprise determining the first score associated with the first operation pattern and the second score associated with the second operation pattern based on at least one evaluation criterion.

[0019] Providing one or more evaluation criteria may increase the accuracy of the determined score. As a result, incorrect ranking of the operation parameter sets may be avoided.

[0020] The medical data may in some examples further comprise physiological parameters associated with the operation parameters and related to the patient.

[0021] 20.136P-WO / 08.08.2025 Many IMDs may be configured to deliver therapy to the patient (e.g., according to at least one operation parameter) and / or measure at least one physiological parameter of the patient (e.g., also according to at least one operation parameter). In other examples, different IMDs may execute said functions and / or an external device may measure the at least one physiological parameter of the patient.

[0022] In general, providing said physiological parameter of the patient may increase the understanding of the patient’s state, the IMD, and the effect it has on the patient (e.g., depending on its operation parameters).

[0023] A second aspect refers to a method for operating an implanted and / or implantable medical device, IMD, wherein the method comprises the steps of: transmitting medical data comprising operation parameters of the IMD; receiving at least one first operation parameter set based at least in part on the medical data; and adjusting an operation configuration of the IMD based at least in part on the at least one first operation parameter set.

[0024] Providing the method allows fine-tuning of the device or system parameters resulting in further improved operation performance.

[0025] The medical data may in some examples further comprise physiological parameters associated with the operation parameters and related to a patient with the IMD.

[0026] A third aspect refers to a method for optimizing operation of an implanted and / or implantable medical device, IMD, wherein the method comprises the steps of: accessing a first operation parameter set related to the operation of the IMD; triggering a test operation with a test operation configuration based at least in part on the at least one first operation parameter set; wherein the test operation configuration varies from the first operation parameter set; and determining an optimized operation parameter set at least in part based on medical data related to the test operation.

[0027] Providing the method allows fine-tuning of the operation parameters resulting in further improved device or system performance. The method according to the third aspect was found to be particularly suitable when it follows the method according to the first aspect of the present invention determining at least one first (and optionally second, third, etc.) operation parameter

[0028] 20.136P-WO / 08.08.2025 set for the IMD of the patient at least in part based on the medical data. Said determined operation parameter sets may pose an advantageous starting point for the method according to the third aspect.

[0029] This concept may generally correspond to applying reinforcement learning.

[0030] The method may, e.g., further comprise determining, for each test operation, an operation score based at least in part on the medical data, wherein determining the optimized operation parameter set may, in some examples, be based at least in part on determining the operation having the highest operation score.

[0031] The method may thus determine how well-suited the operation of the IMD in different test operation configurations is regarding certain criteria that may be related to at least one operation parameter (e.g., battery usage may be minimized) and / or at least one physiological parameter (e.g., a patient-reported and / or differently measured pain score of the patient may be minimized). Determining the operation score may be at least partially based on an action-value function, e.g., a function estimating an expected cumulative future reward for a test operation.

[0032] In some examples, the method may further comprise triggering a further test operation with a further test operation configuration, wherein the further test operation configuration may vary from the optimized operation parameter set in at least one repetition and / or the test operation of the (previously triggered) test operation(s).

[0033] Triggering further test operation(s) may, e.g., be executed repeatedly, e.g., in a loop. Executing said steps repeatedly, i.e., in multiple cycles, may allow to reiteratively improve the IMD operation performance.

[0034] In particular, testing different test operation configurations, i.e., operation of the IMD with operation parameters that, e.g., vary from known (combinations of) operation parameters may be evaluated to gain more insights into how the IMD and the patient (both individually and in their interaction) respond to said variations.

[0035] The medical data may in some examples comprise a plurality of medical data samples. E.g., each medical data sample may comprise at least one operation parameter and at least one

[0036] 20.136P-WO / 08.08.2025 physiological parameter of the patient and the at least one operation parameter may be associated with the at least one physiological parameter.

[0037] Therein, a medical data sample may represent an operation of the IMD, and a patient state of the patient associated with the IMD. Such physiological parameters allow, e.g., modelling the patient in association with their at least one IMD.

[0038] The physiological parameter may, in some examples, comprise at least one of: electroencephalogram data, a heart rate variability, a patient age, a patient height, a patient weight, a patient stimulation log, a prescription drug usage, information on an evoked action compound potential, a patient-reported information, preferably a pain score, and a disability index.

[0039] The operation parameters may, e.g., comprise at least one of: an operation protocol, an operation parameter, and an operation time. In an exemplary, the IMD may be a neurostimulator and the operation parameter may, e.g., comprise at least one of: an electrode polarity, a stimulation amplitude, a stimulation pattern, an electrode identifier, a stimulation frequency, device performance data, and a stimulation pulse duration.

[0040] The inventors found that these examples (for physiological and operation parameters) may be particularly suitable for modelling the patient state and / or the IMD, respectively, in terms of the aspects that may be relevant in the context of the present invention. This may improve the accuracy of the determined optimized operation parameters and / or reduce unwanted side effects that might otherwise be evoked in patients that cannot be modelled as accurately by the physiological parameters described herein.

[0041] A neurostimulator (as an exemplary IMD) typically may comprise a pulse generator and at least one lead. The at least one lead may, e.g., comprise at least one electrode configured to deliver (pulsed) electrical neurostimulation. The at least one electrode may be placed within the patient near the targeted neuron for neurostimulation. The neurostimulation may comprise a series of stimulation pulses characterized by an electrode polarity, a stimulation amplitude, a stimulation pattern, an electrode identifier, a stimulation frequency (e.g., describing the rate at which the stimulation pattern is being repeated), a time between two subsequent pulses within the stimulation pattern, and / or a stimulation pulse duration. All of these may be operation parameters in the context of the present invention.

[0042] 20.136P-WO / 08.08.2025 A fourth aspect refers to an IMD comprising means configured to perform the method as described herein.

[0043] A fifth aspect refers to a data processing device comprising means configured to perform the method as described herein.

[0044] A sixth aspect refers to a computer program comprising instructions, when executed by a computer, causing the computer to perform the method as described herein.

[0045] The IMD according to the fourth aspect, the data processing device according to the fifth aspect, and the computer program according to the sixth aspect provide the advantages described herein in reference to the respective methods.

[0046] It is noted that all methods described herein may include features even if they are rather described with reference to an IMD, processing device, and / or a computer program. Vice versa, aspects that may be described herein with reference to a method may as well be implemented as a corresponding function of an IMD, processing device, and / or a computer program.

[0047] Fig. 1 A Illustration of a patient with an implanted and / or implantable medical device.

[0048] Fig. IB Illustration of a medical device connectivity system comprising a plurality of nodes, i.e., an implanted and / or implantable medical device, a corresponding remote device, a central server, a clinician programmer, and an external database.

[0049] Fig. 2 Illustration of an overview of the methods for optimizing operation according to aspects of an embodiment.

[0050] Fig. 3 Illustration of a first method for operation optimization according to aspects of an embodiment.

[0051] Fig. 4 Illustration of a second method for operation optimization according to aspects of an embodiment.

[0052] 20.136P-WO / 08.08.2025 Fig. 5 Illustration of a method for generating a new IMD according to aspects of an embodiment.

[0053] The following detailed description outlines possible exemplary implementations of the present invention.

[0054] Fig. 1A shows an illustration of a patient 110 with an implanted and / or implantable medical device 100. Therein, the patient 110 may be described by at least one physiological parameter 111 and the IMD 100 may be described by at least one operation parameter 101.

[0055] Fig. IB illustrates a medical device connectivity system (MDCS) 1000 comprising a plurality of nodes, e.g., an IMD 100 of a patient 110 (which may also be denoted as the IMD-patient system 1100), a corresponding remote device 1200, a central server 1300 (e.g., comprising a computing unit 1310 and a central database 1320), a clinician programmer 1400, and an external database 1500.

[0056] In detail, the central server 1300 may serve as a central node in communication with one or more remote devices 1200, one or more clinician programmers 1400, and / or one or more external databases 1500. In other examples, the one or more IMDs 100 may be in direct connection with the central server 1300 - not via the remote device 1200.

[0057] The remote device 1200 may, e.g., be a mobile phone and / or a dedicated patient device. The clinician programmer 1400 may, e.g., be configured to program the IMD 100 remotely via the MDCS 1000 as described herein. The MDCS 1000 in general may allow for communication between its components, e.g., via WiFi, Cellular like 4G and 5G, Bluetooth like Bluetooth Low Energy (BLE), Medical Implant Operation Service (MICS) and / or combinations thereof.

[0058] The methods described herein may, e.g., at least in part be executed by either of the components of the MDCS 1000 described herein.

[0059] Fig. 2 illustrates an overview 200 of methods according to aspects of the present invention. An IMD 100 may be optimized according to two methods 300, 400. A first method 300 (e.g., according to the first aspect) optimizes the IMD 100 based on operation parameters and optionally additionally based on physiological parameters. As a result of the first method 300, operation parameter sets may be determined. A second method 400 (e.g., according to the third

[0060] 20.136P-WO / 08.08.2025 aspect) may further optimize the IMD 100 based on the operation parameter sets determined by the first method 300. The method 300 and method 400 may be combined as one overall method as shown in Fig. 2. The method 300 and method 400 may, however, also be implemented independently from each other.

[0061] Operation data within this disclosure may comprise a set of operation data samples (e.g., 500 samples or more). Each operation data sample may represent an operation related to the IMD 100. An operation related to the IMD may refer to an operation being performed by the IMD, e.g., in the example of the IMD 100 being a SCS delivering neurostimulation to the patient according to one or more operation parameters like neurostimulation amplitude, pattern, frequency, and / or duration. Each operation may be associated with a function executed by the IMD 100. The operation data may help to identify differences between an expected and an actual operation behavior and to identify improvement areas of the operation configuration by reconfiguration (e.g., re-programming the firmware with respect to adjusting operation parameters). Operation data may be logs of performed operations (e.g., providing neurostimulation to the patient, in the example of the IMD being a NS, e.g., an SCS). Operation data may refer to periodically transmitted diagnostic and / or technical data. This may allow to remotely track a status of a patient (e.g., health) or a status of the IMD 100 itself (e.g., whether the IMD operates error free). As a result, a (closed) loop feedback system may be created which allows to continuously analyze and optimize the operation within the IMD 100.

[0062] Fig. 3 illustrates an exemplary method 300 for operation optimization. The method 300 optimizes the operation of an IMD. The method 300 may be performed by a data processingdevice according to aspects of the present invention.

[0063] In step 310, medical data related to at least one operation of the IMD is accessed. Accessing 310 may refer to the receiving the operation data or requesting the medical data from a corresponding data source (e.g., a database).

[0064] In step 320, one or more parameter sets are determined. The operation parameter sets may comprise at least one first operation parameter set. Determining 320 may at least in part based on the medical data accessed in step 310.

[0065] In some implementations, the operation parameter sets may further comprise at least one second operation parameter set. Determining 320 may further comprise the following steps.

[0066] 20.136P-WO / 08.08.2025 In a step 322, at least a first and second operation pattern may be identified. Identifying 322 may at least in part be based on the medical data accessed in step 310. Identifying 322 may for example be based on a location of the IMD or node (e.g., current location or location during operation), a time of day and / or a used operation protocol. Identifying 322 may be done using supervised learning and / or unsupervised learning (e.g., hierarchical clustering, k-means clustering etc.) and / or other suitable data mining / analytic techniques for identifying such patterns (e.g., clusters or policies) in the (unstructured) medical data other similar artificial intelligence-based approaches. For example, unsupervised learning may be based on a machine learning algorithm which leams pattern from unlabeled data. Learning patterns may comprise finding similarities and / or differences and / or pattern and / or structure of the data by itself. On the other hand, supervised learning may be at least partially based on labeled data. For example, supervised learning may comprise training a model based on input objects and desired output objects, e.g., the labels of the input objects. Determining 320 the set of operation parameter sets may include determining the first operation parameter set based on the first operation pattern and determining the second operation parameter set based on the second operation pattern.

[0067] In a first example, a first operation pattern may indicate that operation with a stimulation pattern comprising three stimulation pulses (with a first stimulation amplitude and a first stimulation pulse duration each) per pulse sequence (repeated with a first stimulation frequency) was successful in 100% of the cases while for example a second operation pattern may indicate that operation with a stimulation pattern comprising only one stimulation pulse (with a second stimulation amplitude and a second stimulation pulse duration) repeated with a first stimulation frequency was only successful in 80% of the cases.

[0068] In a second example, it may also be possible that a third operation pattern indicates that duty cycling (e.g., by introducing breaks in which no stimulation is provided to the patients for time periods longer than typical pulse durations and / or pulse sequence durations), for example in the above scenario, may increase battery lifetimes compared to scenarios in which no such duty cycling is used.

[0069] At step 324, the first operation pattern and the second operation pattern are ranked according to a first score associated with the first operation pattern and a second score associated with the second operation pattern. Determining the first and second score (and potentially more

[0070] 20.136P-WO / 08.08.2025 depending on the number of determined operation parameter sets) may be done according to a set of evaluation criteria.

[0071] Said criteria may relate to certain metrics: Metrics positively correlated with pain relief and battery efficiency may contribute to rewards (e.g., high scores). Penalties (e.g., low scores) would be assigned to metrics associated with side effects of the therapy as well as therapy parameters that cannot be achieved due to constraints.

[0072] A first evaluation criteria may be battery usage. For example, a first evaluation criterion may be battery usage. For example, a low battery usage (e.g., 20 mAh / day) may be scored higher (e.g., 1), a medium battery usage (e.g., 100 mAh / day) may be scored medium (e.g., 0.5) and a high battery usage (e.g., 180 mAh / day) may be scored lower (e.g., 0).

[0073] A second evaluation criterion may, e.g., be a pain score of the patient (indicating the pain level experienced by the patient) introduced by the operation. For example, a low pain score may be scored higher (e.g., NRS 1), a medium pain score may be scored medium (e.g., NRS 5) and a high pain score may be scored lower (e.g., NRS 9).

[0074] Evaluation criteria may have different weights (i.e., one criterion being more important than another criteria). Accordingly, an evaluation criterion having a higher weighting (e.g., primary weighting) may have a larger impact on a (total) score than an evaluation criterion having a lower weighting (e.g., secondary weighting).

[0075] Applying the one or more evaluation criteria to the identified operation patterns may result in the determined score associated with each operation pattern. Based on the score, the operation patterns may be ranked (e.g., in descending order based on the associated score). In examples above, the first operation pattern may be scored lower than the second operation pattern.

[0076] At step 326, a preference of the first operation parameter set and the second operation parameter set may be defined according to the ranking of step 324. The preference may indicate which operation parameter set is to be used under which circumstances.

[0077] In step 330, the at least one first (and optionally the second, third, etc.) operation parameter set is transmitted to the IMD.

[0078] 20.136P-WO / 08.08.2025 The present invention further relates to a corresponding method for operating an IMD based on an operation parameter set. The method may at least partly be performed by the IMD. The operation parameter set may be determined according to method 300. The method for operating the node may comprise the following steps.

[0079] First, operation data related to the at least one operation path may be transmitted to a data- processing device.

[0080] Second, a set of operation parameter sets comprising at least one first operation parameter set may be received from the data-processing device.

[0081] Third, an operation configuration of the node is adjusted based at least in part on the set of operation parameter sets. Adjusting the operation configuration may comprise configuring the IMD such that the IMD uses a first operation parameter set as a primary solution (i.e., a default configuration) and a second operation parameter set as a secondary solution (i.e., in case the primary solution is not available / applicable use the second operation parameter set as a fallback configuration). There may be a primary and secondary solution for first circumstances and another primary and secondary solution for second circumstances (i.e., different from the first circumstances). Adjusting the operation configuration may in addition or alternatively comprise operating the IMD according to the operation configuration.

[0082] Fig. 4 illustrates a method 400 for optimizing operation of an IMD according to aspects of the present invention. The method 400 may, e.g., be performed by a data-processing device according to aspects of the present invention.

[0083] In step 410, an operation parameter set related to the IMD is accessed. Accessing 410 may refer to the receiving the set of operation parameter sets (and optionally physiological data) (e.g., from the IMD) or requesting the operation parameter set (and optionally physiological data) from a corresponding data source (e.g., a database). The optimized operation parameter sets may be determined using method 300 of the present invention, for example.

[0084] In step 420, at least one operation with a test configuration for at least one operation parameter set of the set of operation parameter sets may be triggered. A test configuration may indicate an adjustment of at least one operation parameter of the operation parameter set. The adjustment may be, e.g., [frequency] -related. The adjustment may indicate to update a corresponding

[0085] 20.136P-WO / 08.08.2025 operation parameter by a defined amount. Updating may comprise reducing the operation parameter by the defined amount (e.g., an operation parameter-defined amount) or increasing the operation parameter by the defined amount (e.g., an operation parameter-defined amount). Updating may also comprise selecting one option of a plurality of options for the given operation parameter (e.g., selecting an operation protocol out of a plurality of available operation protocols).

[0086] The adjustment may, e.g., comprise reducing a stimulation amplitude by a defined amount. The reduction of the stimulation amplitude may for example be 3 mA, 2.5 mA, 2 mA, etc. Additionally, or alternatively, the adjustment may, e.g., comprise reducing a stimulation frequency by a defined amount. The reduction of the stimulation frequency may for example be 600 Hz, 400Hz, 300 Hz, etc. Additionally, or alternatively, the adjustment may, e.g., comprise reducing a stimulation pulse duration by a defined amount. The reduction of the stimulation duration may for example be 300 ms, 250 ms, 200 ms, etc. The above examples may, e.g., apply in an IMD being a neurostimulator, e.g., an SCS and may be directed at increasing the battery lifetime of the IMD, optionally by sustaining the patient-experienced (chronic) pain at the same time.

[0087] In step 430, an optimized operation parameter set may be determined. Determining 430 may be at least in part based on operation data related to the at least one (triggered) operation. The optimized operation parameter set may be feedback (e.g., transmitted to a central sever) and used for (parts of) upcoming IMDs (e.g., IMDs which are to be manufactured). This way a manufacturing process of IMDs may be improved as the optimized operation parameter set may give insights about performance metrics between corresponding operation parameters (e.g., optimum tradeoff between battery usage, stimulation amplitude, stimulation frequency, stimulation pattern, and / or stimulation pulse duration).

[0088] In some implementations the method 400 may comprise the optional step 422. In step 422, for each of the at least one (triggered) operation, an operation score may be determined based on the operation data. The operation data may be accessed. The operation score may for example be determined based on at least one physiological parameter (e.g., a patient’s pain score) and / or at least one operation parameter (e.g., battery usage rate). Accordingly, operation with a low battery usage and / or a low pain score may result in a higher operation score than an operation with a high battery usage and / or a high pain score.

[0089] 20.136P-WO / 08.08.2025 Determining 430 may further be based on determining the operation having the highest operation score. Determining the test configuration (i.e., the adjustment for the at least one operation parameter) may be predefined by a standardized test configuration protocol. Alternatively, the test configuration may be dynamically determined. Dynamically determining may comprise iteratively adjusting at least one operation parameter, triggering a corresponding operation, and determining based on the operation score how to define the test configuration for a following iteration. This way, an operation parameter set having the highest operation score may be determined resulting in the optimized operation parameter set. Determining the optimized operation parameter set as described with respect to method 400 may be implemented using reinforcement learning or other suitable optimization techniques (e.g., generic algorithms like particle swarm optimizer).

[0090] The present invention further relates to a corresponding method for optimizing operation of an IMD. The method may be performed by a node of the MDCS (cf. Fig. IB) or the IMD. The method may comprise the following steps.

[0091] First, receiving a trigger for performing for at least one operation parameter set at least one operation with a test configuration. The received trigger may correspond to the triggering 420 of method 400. Performing the operation for the at least one operation parameter may comprise adjusting the operation parameter according to the test configuration.

[0092] Second, performing the at least one operation for an operation parameter set may comprise adjusting an operation parameter of the operation parameter set according to the test configuration. Operation data of the performed operation may be transmitted to a data-processing device from which the trigger was received or stored into a corresponding data source.

[0093] Third, receiving an optimized operation parameter set at least in part based on operation data related to the at least one (performed) operation. The optimized operation parameter set may be received from the data-processing device.

[0094] Fig. 5 illustrates a method 500 for generating a new implantable medical device or medical device connectivity system according to aspects of the present invention. The method 500 may be performed by a data-processing device like a server having access to one or more IMDs 100. The method 500 may be seen as a potential add-on to method 400. While methods 300 and 400, as shown for example with respect to Fig. 2, may provide optimization for one IMD 100, method

[0095] 20.136P-WO / 08.08.2025 500 may generally be used to generate a new IMD or define the relevant aspects of an associated medical device connectivity system (i.e. it considers all components of the MDCS as per its operation as a network). Generating may for example refer to determining a construction design for the new IMD. The new IMD may refer to a future version of the current / previous version of the IMD 100. The method 500 may use data (e.g., the optimized operation parameter set) of one or more IMDs 100 to determine an initial operation parameter set, which performs better in more operating environments than the initial operation parameter set of previous MDCSs. The method 500 may comprise the following steps.

[0096] In step 510, the optimized operation parameter set(s) of one or more IMDs 100 are accessed. This may be done by requesting and receiving the optimized operation parameter set from each of the one or more IMDs 100. Alternatively, this may be done by requesting the one or more optimized operation parameter set(s) from a data source (e.g., a data base). Accessing may be done over a public or private network used by the one or more IMDs 100. Apart from the optimized operation parameter set, corresponding operation data and operation patterns of the one or more IMDs 100 may optionally be accessed.

[0097] In step 520, an initial operation parameter set is determined based on the one or more optimized operation parameter set(s). Determining may be done using a machine learning model. Depending on a selected parameter set of the optimized operation parameter set(s), the model may comprise a regression model, a classification model, or a combination thereof. The selected parameter set may refer to a subset of operation parameters (e.g., the initial operation parameter set may comprise some operation parameters which are adjusted and some which are kept the same with respect to the previous initial operation parameter set). The model may be trained using the one or more optimized operation parameter set(s) and / or the corresponding operation data and operation patterns. As a result, the model may be able to predict for a given input (e.g., a candidate for an initial operation parameter set) a corresponding output (e.g., performance data of operation performed using the candidate for an initial operation parameter set). This way the initial operation parameter may be determined by inputting a one or more candidates into the model, evaluating the corresponding output (e.g., with respect to performance data of operation performed using the one or more candidates) and selecting the best candidate (e.g., the candidate achieving the best performance data) as the initial operation parameter set.

[0098] In step 530, a new IMD may be generated at least based on the initial operation parameter set. Generating may comprise determining a construction design for the new IMD. The construction

[0099] 20.136P-WO / 08.08.2025 design may comprise an initial operation configuration determined based on the initial operation parameter set. The construction design may further comprise additional or new hardware (i.e., a hardware upgrade). This may be the case if the initial operation parameter set requires the hardware upgrade in order to actually be configurable (e.g., a determined initial signal strength of the initial operation parameter set may only be configurable if a corresponding new hardware feature is added into the new IMD. This way, design, or construction decisions for future / new IMDs may be improved.

[0100] Generally, the method 500 for generating a new implantable medical device or medical device connectivity system may also be applied to medical and / or electronical devices associated with the IMD and / or general medical devices, e.g., the method 500 may be used to generate new medical devices associated with the IMD and / or medical devices in general. For example, the IMD may be associated with a patient remote device, e.g., a smartphone and / or a tablet computer and / or a home station of the IMD. Particularly, applying the method 500 with respect to the patient remote device may comprise generating a new patient remote device. For example, generating a new patient remote device may be at least partially based on generating a new IMD, e.g., generating a new IMD and generating a new patient remote device may comprise an alternating optimization process. In addition, or alternatively, the IMD may be associated with a clinician user interface and / or a remote monitoring server and / or a medical data base.

[0101] The methods according to the present invention may be implemented in terms of a computer program which may be executed on any suitable data processing device comprising means (e.g., a memory and one or more processors operatively coupled to the memory) being configured accordingly. The computer program may be stored as computer-executable instructions on a non-transitory computer-readable medium.

[0102] Embodiments of the present disclosure may be realized in any of various forms. For example, in some embodiments, the present invention may be realized as a computer-implemented method, a computer-readable memory medium, or a computer system.

[0103] In some embodiments, a non-transitory computer-readable memory medium may be configured so that it stores program instructions and / or data, where the program instructions, if executed by a computer system, cause the computer system to perform a method, e.g., any of the method embodiments described herein, or, any combination of the method embodiments described

[0104] 20.136P-WO / 08.08.2025 herein, or, any subset of any of the method embodiments described herein, or, any combination of such subsets.

[0105] In some embodiments, a computing device may be configured to include a processor (or a set of processors) and a memory medium, where the memory medium stores program instructions, where the processor is configured to read and execute the program instructions from the memory medium, where the program instructions are executable to implement any of the various method embodiments described herein (or, any combination of the method embodiments described herein, or, any subset of any of the method embodiments described herein, or, any combination of such subsets). The device may be realized in any of various forms.

[0106] Although specific embodiments have been described above, these embodiments are not intended to limit the scope of the present disclosure, even where only a single embodiment is described with respect to a particular feature. Examples of features provided in the disclosure are intended to be illustrative rather than restrictive unless stated otherwise. The above description is intended to cover such alternatives, modifications, and equivalents as would be apparent to a person skilled in the art having the benefit of this disclosure.

[0107] The scope of the present disclosure includes any feature or combination of features disclosed herein (either explicitly or implicitly), or any generalization thereof, whether or not it mitigates any or all of the problems addressed herein. In particular, with reference to the appended claims, features from dependent claims may be combined with those of the independent claims and features from respective independent claims may be combined in any appropriate manner and not merely in the specific combinations enumerated in the appended claims.

[0108] 20.136P-WO / 08.08.2025

Claims

Claims1. Method for optimizing operation of an implanted and / or implantable medical device, IMD (100), the method comprising: accessing medical data comprising operation parameters (101) of a plurality of IMDs of a plurality of patients (110); determining at least one first operation parameter set for at least one IMD of at least one patient at least in part based on the medical data; and transmitting the at least one first operation parameter set to the at least one IMD.

2. The method of claim 1, further comprising determining at least one second operation parameter set for the at least one IMD (100) of the at least one patient (110) at least in part based on the medical data, wherein the determining of the first and second operation parameter sets comprises: identifying at least a first and a second operation pattern at least in part based on the medical data; determining the first operation parameter set based on the first operation pattern and determining the second operation parameter set based on the second operation pattern; ranking the first operation pattern and the second operation pattern according to a first score associated with the first operation pattern and a second score associated with the second operation pattern; and defining a preference of the first operation parameter set and the second operation parameter set according to the ranking.

3. The method of claim 2, further comprising: determining the first score associated with the first operation pattern and the second score associated with the second operation pattern based on at least one evaluation criterion.

4. The method of any of claims 1 to 3, wherein the medical data further comprise physiological parameters (111) associated with the operation parameters (101) of at least one of the plurality of IMDs (100) of at least one of the plurality of patients (110).

5. Method for operating an implanted and / or implantable medical device, IMD (100), wherein the method comprises: transmitting medical data comprising operation parameters (101) of the IMD;20.136P-WO / 08.08.2025receiving at least one first operation parameter set based at least in part on the medical data; and adjusting an operation configuration of the IMD based at least in part on the at least one first operation parameter set.

6. The method of claim 5, wherein the medical data further comprise physiological parameters (111) associated with the operation parameters (101) and related to a patient (110) with the IMD (100).

7. Method for optimizing operation of an implanted and / or implantable medical device, IMD (100), wherein the method comprises: accessing a first operation parameter set related to the operation of the IMD; triggering a test operation with a test operation configuration based at least in part on the at least one first operation parameter set; wherein the test operation configuration varies from the first operation parameter set; and determining an optimized operation parameter set at least in part based on medical data related to the test operation.

8. The method of claim 7, wherein the method further comprises: determining, for each test operation, an operation score based at least in part on the medical data; and wherein determining the optimized operation parameter set is further based at least in part on determining an operation having the highest operation score.

9. The method of claim 7 or 8, further comprising triggering a further test operation with a further test operation configuration, wherein the further test operation configuration varies from the optimized operation parameter set.

10. The method of any of claims 1 to 9, wherein the medical data comprises a plurality of medical data samples, wherein each medical data sample comprises at least one operation parameter (101) and at least one physiological parameter (111) of the patient (110); wherein the at least one operation parameter is associated with the at least one physiological parameter.20.136P-WO / 08.08.202511. The method of claim 4, 6 or 10, wherein the physiological parameter (111) comprises at least one of: electroencephalogram data, a heart rate variability, a patient age, a patient height, a patient weight, a patient stimulation log, a prescription drug usage, information on an evoked action compound potential, a patient-reported information, preferably a pain score, and a disability index.

12. The method of any of claims 1 to 11, wherein the operation parameter (101) comprises at least one of: an operation protocol, and an operation time; and wherein optionally the IMD is a neurostimulator and the operation parameter comprises at least one of: an electrode polarity, a stimulation amplitude, a stimulation pattern, an electrode identifier, a stimulation frequency, device performance data, and a stimulation pulse duration.

13. An implanted and / or implantable medical device, IMD (100), comprising means configured to perform the method of any of claims 1 to 12.

14. A data processing device comprising means configured to perform the method of any of claims 1 to 12.

15. A computer program comprising instructions for performing the method of any one of claims 1 to 12.20.136P-WO / 08.08.2025

Images