Device and system for managing a knowledge database for therapy selection

Abstract

A computer-readable storage medium containing instructions that cause at least one processor to implement a computer-implemented procedure for managing a knowledge database (11) for therapy selection when the instructions are executed by the at least one processor, the procedure comprising the following steps: - Retrieving a data structure (D) from the knowledge database (11), wherein the data structure (D) has at least one therapy option (T1, T2, T3, T4) for a disease depending on at least one decision criterion (e1); - Recording case information (I); - Evaluating at least one decision criterion (e1) based on the case information (I); - Generating a selection therapy proposal (T) based on the data structure (D) and the case information (I); - Recording evaluation information (E) regarding the selection therapy proposal (T); - Generating an updated data structure (D') based on the evaluation information (E); and - Transferring the updated data structure (D') to the knowledge database (11).

Description

[0001] The invention relates to a method for managing a knowledge database for therapy selection, particularly for chronic inflammatory bowel diseases. Furthermore, the invention relates to a device and a system for implementing the method.

[0002] Knowledge databases serve to store knowledge in a structured format, typically so that this knowledge can be made accessible to a wider audience. Treating physicians can also access information from knowledge databases when treating diseases, such as chronic inflammatory bowel disease (IBD). This information comes, for example, from guidelines for treating a patient with a specific condition or current publications. Such a guideline can, for example, specify a patient's drug therapy, with the dosage specified depending on the patient's body weight.

[0003] A common problem with IBD treatment guidelines is that they are formulated too vaguely for the treatment of individual patients. The course of IBD is typically very individual. This often results in a lack of concrete recommendations for everyday medical practice.

[0004] Current IBD treatment is based on a combination of traditional knowledge, guidelines, and individual treatment decisions. Individual treatment decisions, if documented at all, are often not documented in a structured format, making them difficult to locate or understand.

[0005] For these reasons, there is a need to specify a method for managing a knowledge database for therapy selection. In particular, the method should enable the inclusion of individual therapy decisions—in addition to traditional knowledge and guidelines—in the knowledge database.

[0006] The task is solved, in particular, by a method for managing a knowledge database for therapy selection. The method comprises the following steps: - Retrieving a data structure from the knowledge database, wherein the data structure has at least one therapy option for a clinical picture depending on at least one decision criterion; - Collecting case information; - Evaluating at least one decision criterion based on the case information; - Generating a selection therapy proposal based on the data structure and case information; - Collecting evaluation information regarding the selection therapy proposal; - Generating an updated data structure based on the evaluation information; and - Transferring the updated data structure to the knowledge base.

[0007] The described procedure is based on data structures from a knowledge database. These data structures specify at least one treatment option for a specific clinical picture, depending on a decision criterion. A decision criterion can be understood as a condition or a logical combination of several conditions under which a particular treatment option is recommended or indicated. For example, a treatment option might consist of administering a specific medication under the condition that the patient is suffering from a severe flare-up and has a body weight of at least 50 kg.

[0008] The decision criterion can be evaluated based on case information that provides information about a specific patient and / or the patient's disease activity (e.g., remission, activity of the underlying disease, etc.). In the above example, case information on the severity of the disease flare and body weight provide the information needed to evaluate whether the specified treatment option can be recommended. The case information can be captured, for example, through user input. Alternatively, the case information can be imported into the knowledge database in digital form. In particular, the case information can have been previously entered into a practice management system that is linked to the knowledge database.

[0009] After collecting case information and subsequently evaluating at least one decision criterion, the described procedure generates a treatment suggestion that can be interpreted as a concrete recommendation for action for the treating physician. The treatment suggestion can be a single treatment option. Furthermore, the treatment suggestion can be a sequence of multiple treatment options and / or courses of action.

[0010] The selection therapy proposal can be generated algorithmically, in particular by evaluating decision criteria based on case information, and / or manually, i.e. depending on user input.

[0011] A core concept of the present invention is that evaluation information is recorded regarding the selected therapy suggestion, linked to the therapy suggestion, and transferred back to the knowledge database as an updated data structure. The evaluation information can, for example, indicate that a corresponding selected therapy suggestion has been used for the patient's therapy. Thus, information about the frequency of use of the respective therapy options is collected in the knowledge database. Furthermore, further evaluation information, such as "positive" or "negative" ratings, additional comments, etc., can be recorded for the selected therapy suggestion.

[0012] This allows a knowledge database to be managed dynamically rather than as a static source of information, essentially as a kind of ongoing study whose results are accessible at any time. Based on such a knowledge database, therapy selection can be optimized. In particular, the knowledge database allows for an integrative procedure that uses parameters routinely recorded at different times and uses scaling based on scores to capture disease activity and derive therapy adjustments (e.g., therapy intensification, therapy change, etc.).

[0013] In one embodiment, the data structure is a graph with a set of nodes and a set of edges. The set of nodes comprises two types of nodes: decision nodes and clue nodes. The decision nodes correspond to the at least one decision criterion, and the clue nodes contain information related to the at least one treatment option.

[0014] This embodiment has the advantage that graphs represent flexible data structures that can be processed efficiently by computers.

[0015] In one embodiment, generating the selected therapy suggestion comprises determining a path in the graph. Determining the path may, in particular, include evaluating the at least one decision criterion. The path may be determined such that a subsequent node of a decision node in the path is determined according to the evaluation result of the at least one decision criterion.

[0016] In one embodiment, generating the selection therapy suggestion comprises calculating a numerical value based on the data structure using the case information. In particular, the numerical value can be a medication dose, which is calculated, for example, based on a patient's body weight. The body weight can be stored in the case information, and the data structure can contain a (generic) therapy option that describes medication administration depending on the patient's body weight (e.g., "1 mg cortisone per 1 kg body weight").

[0017] In one embodiment, all edges of the edge set, or at least the edges of a subset thereof, are assigned a weighting value. In particular, the weighting value can be determined based on evaluation information that was acquired in a previous execution of the method and transferred to the database. For example, a high weighting value can express that a therapy option corresponding to the edge was frequently used for the patient's therapy. Conversely, a low weighting value can express that a therapy option corresponding to the edge was rarely selected for the patient's therapy.

[0018] Instead of the frequency of use in therapy, the weighting value can also indicate an evaluation value, where a high evaluation value can be considered as a positive evaluation and a low evaluation value as a negative evaluation.

[0019] Instead of a single weighting value, a vector of weighting values ​​can be provided to capture several different aspects (e.g. frequency of use, rating value, etc.).

[0020] This embodiment provides the therapy options in the knowledge database with additional information that can be included in the selection of a suitable therapy suggestion.

[0021] In one embodiment, the method further comprises visualizing the data structure as a decision graph. For example, decision nodes can be represented in the form of a diamond, clue nodes in the form of a rectangle, and edges in the form of linear connections between the nodes. In particular, the strength of an edge can depend on its weight, preferably proportionally. Furthermore, nodes, edges, or paths can be represented in color, for example, to distinguish positive events from negative events.

[0022] This embodiment provides the user with a visualization of the data structure containing the at least one therapy option and the at least one decision criterion. This can facilitate the user's selection of a therapy suggestion. Furthermore, the user can obtain an overview of various therapy options, their interrelationships, and the associated information. In particular, a representation in which the strength of an edge is selected proportional to its weight contributes to better orientation of the user in the decision graph.

[0023] The object is further achieved by a computer-readable storage medium. The computer-readable storage medium contains instructions that cause at least one processor to implement the method as described above when the instructions are executed by the at least one processor.

[0024] The computer-readable storage medium offers similar advantages and technical effects as those described in connection with the method.

[0025] The object is further achieved by a device for managing a knowledge database for therapy selection. The device can, in particular, comprise the following: - a receiving unit configured to receive a data structure from the knowledge database, wherein the data structure comprises at least one therapy option for a clinical picture depending on at least one decision criterion; - a processing unit configured to perform the following steps: - Collecting case information; - Evaluating at least one decision criterion based on the case information; - Generating a selection therapy proposal based on the data structure and case information; - Collecting evaluation information regarding the selection therapy proposal; - Generating an updated data structure based on the evaluation information; - a transmitting unit designed to transfer the updated data structure to the knowledge database.

[0026] The device offers similar advantages and technical effects as those described in connection with the method.

[0027] The problem is further solved by a system for managing a knowledge database for therapy selection. The system may, in particular, comprise the following: - a server having a knowledge database and a server communication interface; and - at least one client unit comprising a device as described above.

[0028] The server is designed to provide at least one data structure from the knowledge database via the server communication interface upon request from the at least one client unit.

[0029] Further embodiments emerge from the subclaims.

[0030] The invention is described below using exemplary embodiments, which are explained in more detail with reference to the figures. Herein: Fig. 1 a flowchart of a method according to the invention according to an embodiment; Fig. 2 a visualization of a data structure according to a first embodiment; Fig. 3 shows a section of a visualization of a data structure according to a second embodiment; Fig. 4 an inventive system according to an embodiment.

[0031] In the following description, the same reference numbers are used for identical and equivalent parts.

[0032] Fig. Figure 1 schematically shows the flow of a method according to the invention for managing a knowledge database for therapy selection. According to this exemplary embodiment, a therapy suggestion for a patient suffering from a chronic inflammatory bowel disease is to be determined, and evaluation information about the treatment carried out according to the therapy suggestion is to be transferred back to the knowledge database.

[0033] In step S1, the data structure D is retrieved from the knowledge database 11. The data structure D contains a variety of treatment options for a clinical picture, for example, an IBD clinical picture, depending on various decision criteria.

[0034] In step S2, case information I is collected, which provides relevant information about the patient (e.g., age, gender, body weight, previous illnesses, etc.) and their medical history (e.g., laboratory data, treatments performed, disease flare-ups). This case information enables decision criteria regarding treatment options from data structure D to be answered.

[0035] Subsequently, in step S3, the decision criteria are evaluated based on the case information and the selection therapy proposal T is generated for the patient. These steps are exemplified in the context of Fig. 2 explained.

[0036] In step S4, evaluation information E is collected through user input, which indicates, for example, that the patient was treated according to the determined selection therapy suggestion T and that the selection therapy suggestion T was evaluated positively.

[0037] In step S5, the data structure is updated based on the evaluation information E, for example, by adjusting the numerical values ​​indicating the number of uses and the number of positive evaluations of the selected therapy suggestion T. In this way, an updated data structure D' is generated.

[0038] Finally, the updated data structure D' is transferred back to the knowledge database 11 in step S6.

[0039] Fig. Figure 2 shows a visualization of a data structure related to the clinical problem case “treatment failure under TNF-alpha therapy”.

[0040] In the embodiment shown, the decision criterion e1 is whether the patient responds to TNF-alpha therapy, or responds only inadequately or not at all.

[0041] In case of remission, i.e. successful therapy, no further measures are required (right path starting from decision criterion e1).

[0042] In cases of inadequate response (left path starting from decision criterion e1), further recommended procedures and courses of action are presented, such as TNF-α determination. Depending on the measured TNF-α level, the data structure shown contains further treatment options T1 and T2. In particular, treatment option T1 "switch anti-TNF or additional AZA, 6-MP" is recommended for patients with good anti-TNF levels, and treatment option T2 "dose intensification and / or interval shortening" is recommended for patients with low anti-TNF levels.

[0043] In particular, assuming that the case information I can be used to decide the decision criterion e1 (ie whether there has been a sufficient response to anti-TNF therapy) and the anti-TNF level can be assessed (good or too low), one of the treatment options T1 or T2 can be chosen as a selection therapy proposal.

[0044] Furthermore, Fig. 2 the information field 30 is shown, which can contain additional information (e.g. literature references).

[0045] Fig. Figure 3 shows an excerpt from a visualization of a data structure according to a second embodiment. In this embodiment, therapy option T3 "Mesalazine therapy" and therapy option T4 "Budesonide MMX 9mg" are proposed for the clinical picture "mild disease." The strengths of edges k3 and k4, which lead to the respective therapy options T3 and T4, are different because edges k3 and k4 have different weights in the data structure. This is due to the fact that a higher number of uses was registered for therapy option T4 via the recorded evaluation information. In comparison, edge k3 is less pronounced, which indicates a lower frequency of use of therapy option T3.

[0046] Fig. 4 shows the structure of an inventive system for managing the knowledge database 11, according to an embodiment.

[0047] The system comprises a server 10 with the knowledge database 11 and a server communication interface 12. The server 10 can be located, for example, in a data center.

[0048] The system further comprises a device 20 for managing the knowledge database 11, which is provided on a client unit. In particular, the client unit can be located in a doctor's office, a hospital, etc., and connected to the server 10 via a network.

[0049] Upon request from the client unit, the server 10 provides the device 20 with the data structure D. The request may, for example, contain a clinical picture or a problem case (e.g., "treatment failure under TNF-alpha therapy").

[0050] The device 20 has a receiving unit 21, a processing unit 22 and a transmitting unit 23. The data structure D is received from the server communication interface 12 via the receiving unit 21. The processing unit 22 is designed to implement the method steps of the described method (for example, as described in connection with Fig. 1). The updated data structure D' can be sent back to the server communication interface 12 via the transmitting unit 23.

[0051] In other words, the system is configured to manage a knowledge database 11 by the device 20 obtaining a part of the knowledge database 11 (ie the data structure D) from the server 10, updating it locally (in particular by collecting evaluation information) and sending it back to the server 10.

[0052] It is self-evident that the system can be extended in an obvious manner to several client units, with the knowledge database 11 being provided centrally by the server 10.

[0053] At this point it should be noted that all of the parts described above are to be regarded individually - even without additional features described in the respective context, even if these have not been explicitly identified as optional features in the respective context, e.g. by using: in particular, preferably, for example, e.g., if necessary, round brackets, etc. - and in combination or any sub-combination as independent embodiments or further developments of the invention, as defined in particular in the introduction to the description and the claims. Deviations from this are possible. Specifically, it should be noted that the word in particular or round brackets do not identify any features that are mandatory in the respective context. List of reference symbols S1 Retrieving the data structure from the knowledge base S2 Collecting case information S3 Evaluating a decision criterion and generating the selection therapy proposal S4 Collecting evaluation information S5 Updating the data structure S6 Transfer the updated data structure to the knowledge database D data structure D' updated data structure E Evaluation information e1 Decision criterion I Case information k3, k4 edge T Selection therapy suggestion T1, ..., T4 treatment options 11 Knowledge base 12 Server communication interface 20 Device 21 Receiving unit 22 processing unit 23 Transmitter unit 30 Information field

Claims

[1] A computer-readable storage medium containing instructions that cause at least one processor to implement a computer-implemented method for managing a knowledge database (11) for therapy selection when the instructions are executed by the at least one processor, the method comprising the following steps: - retrieving a data structure (D) from the knowledge database (11), wherein the data structure (D) has at least one therapy option (T1, T2, T3, T4) for a clinical picture depending on at least one decision criterion (e1); - Collecting case information (I); - evaluating at least one decision criterion (e1) on the basis of the case information (I); - Generating a selection therapy proposal (T) based on the data structure (D) and the case information (I); - Collecting evaluation information (E) regarding the selection therapy proposal (T); - generating an updated data structure (D') based on the evaluation information (E); and - Transferring the updated data structure (D') to the knowledge database (11). [2] Computer-readable storage medium according to claim 1, wherein the data structure (D) is a graph with a node set and an edge set, wherein the node set comprises decision nodes and hint nodes, wherein the decision nodes correspond to the at least one decision criterion (e1) and the hint nodes have information relating to the at least one therapy option (T1, T2, T3, T4). [3] Computer-readable storage medium according to one of the preceding claims, in particular according to claim 2, wherein the generation of the selection therapy proposal (T) comprises determining a path in the graph, wherein in particular a successor node of a decision node in the path is determined according to the evaluation of the at least one decision criterion (e1). [4] Computer-readable storage medium according to one of the preceding claims, wherein generating the selection therapy proposal (T) comprises calculating a numerical quantity, in particular a drug dose, on the basis of the data structure (D), using the case information (I). [5] Computer-readable storage medium according to one of the preceding claims, in particular according to one of claims 2-3, wherein the edges (k3, k4) of at least a subset of the edge set are assigned a weighting value which is determined in particular based on previous evaluation information. [6] Computer-readable storage medium according to one of the preceding claims, in particular according to claim 5, further comprising visualizing the data structure (D) as a decision graph, wherein the strength of an edge (k3, k4) is dependent on its weight, in particular proportional. [7] Device (20) for managing a knowledge database (11) for therapy selection, the device (20) comprising the following: - a receiving unit (21) designed to receive a data structure (D) from the knowledge database (11), wherein the data structure (D) has at least one therapy option (T1, T2, T3, T4) for a clinical picture depending on at least one decision criterion (e1); - a processing unit (22) designed to perform the following steps: - Collecting case information (I); - evaluating at least one decision criterion (e1) on the basis of the case information (I); - Generating a selection therapy proposal (T) based on the data structure (D) and the case information (I); - Collecting evaluation information (E) regarding the selection therapy proposal (T); - generating an updated data structure (D') based on the evaluation information (E); - a transmitting unit (23) which is designed to transmit the updated data structure (D') to the knowledge database (11). [8] System for managing a knowledge database for therapy selection, the system comprising the following: - a server (10) having a knowledge database (11) and a server communication interface (12); and - at least one client unit having a device (20) according to claim 7, wherein the server (10) is designed to provide at least one data structure (D) from the knowledge database (11) via the server communication interface (12) upon request from the at least one client unit.

Images