Method for generating inference models, server equipment, computer programs, and methods for supporting information provision.

Abstract

To increase motivation for providing support information to patients. [Solution] Prescription information PR related to the prescription of medication to a patient is transmitted from the client device 11 to the server device 12. Based on the prescription information PR, the server device 12 infers an index value related to at least one of the implementation and effectiveness of providing support information to the patient. Inference result information RS associated with the index value is transmitted from the server device 12 to the client device 11. The client device 11 is provided with the inference result information RS.

Description

Technical Field

[0001] The present disclosure relates to a method for generating an inference model by an arithmetic device. The present disclosure also relates to a server device capable of communicating with a client device. The present disclosure also relates to a computer program executable by a processor mounted on the server device. The present disclosure also relates to a method for assisting in providing information to a patient executed by the server device and the client device.

Background Art

[0002] Patent Document 1 discloses a system that infers guidance content to be provided to a patient by inputting information related to the health state of the patient and information related to drugs prescribed to the patient.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] There is a need to enhance the motivation for providing support information to patients.

Means for Solving the Problems

[0005] One exemplary aspect that can be provided by the present disclosure is a method for generating an inference model by an arithmetic device, comprising: preparing learning data in which feature values extracted from support information provided to a patient are associated with index values related to at least one of the implementation and effect of the provision of the support information; performing machine learning using the learning data so as to construct an inference model that extracts the feature values from prescription information related to a drug prescription for an input patient and infers the index values.

[0006] One example of an embodiment that may be provided by this disclosure is a server device capable of communicating with a client device, An interface for receiving prescription information related to the prescription of medications to patients from the client device, A processor that infers an index value related to at least one of the implementation and effectiveness of providing support information to the patient based on the prescription information, and causes the client device to provide the inference result information associated with the index value, It is equipped with.

[0007] One example of an embodiment that may be provided by this disclosure is a computer program executable by a processor installed in a server device capable of communicating with a client device, By being executed, the server device will Prescription information related to the prescription of medication to patients is received from the client device. Based on the aforementioned prescription information, infer an indicator value related to at least one of the implementation and effectiveness of providing support information to the patient, The client device is provided with inference result information associated with the aforementioned index value.

[0008] One example of an embodiment that may be provided by this disclosure is a method for supporting the provision of information to a patient, which is performed by a client device and a server device, Prescription information relating to the prescription of medication for the aforementioned patient is transmitted from the client device to the server device. In the server device, based on the prescription information, an index value relating to at least one of the implementation and effectiveness of providing support information to the patient is inferred, The server device transmits the inference result information associated with the aforementioned index value to the client device. The client device provides the inference result information.

[0009] According to the configurations described in each of the above examples, the user of the client device can obtain information relating to at least one of the implementation and effectiveness of the provision of support information related to a prescription by inputting prescription information relating to a drug prescribed to a patient, prior to its implementation. This can increase the motivation of healthcare professionals to provide support information to patients.

[0010] Providing such supportive information can encourage patients to become more proactive in their treatment, thereby promoting repeat visits to medical institutions. This can not only improve the patient's health but also improve the financial performance of the medical institution. Improved medication adherence among patients requiring treatment can also improve the financial performance of pharmaceutical companies.

[0011] For service providers offering support information, acquiring information related to at least one of the implementation and effectiveness of that support information allows them to identify highly useful support information. This, in turn, encourages the creation and provision of even higher-quality support information. [Brief explanation of the drawing]

[0012] [Figure 1] This illustrates the configuration of a support system according to one embodiment. [Figure 2] Figure 1 illustrates the functional configuration of the client and server devices. [Figure 3] Figure 1 illustrates the functional configuration of the inference model generation device. [Figure 4] Figure 3 illustrates the processing flow executed by the processor. [Figure 5] Figure 2 shows an example of the screen displayed on the client device. [Figure 6] Figure 2 shows another example of the screen displayed on the client device. [Modes for carrying out the invention]

[0013] While referring to the accompanying drawings, examples of embodiments will be described in detail below. In each of the drawings referred to in the following description, the scale is changed as necessary to make each element illustrated recognizable size. The scale is changed.

[0014] As used herein, the term "medical institution" means an institution where medical staff provides explanations about pharmaceuticals to patients. Examples of "medical institutions" include hospitals, clinics, clinics, dispensaries, pharmacies, drugstores, etc. Examples of "medical staff" include doctors, nurses, pharmacists, care managers, dietitians, etc.

[0015] FIG. 1 illustrates the configuration of a support system 10 according to an example of an embodiment. The support system 10 includes a client device 11 and a server device 12.

[0016] The client device 11 is a device that provides a user with health advice for supporting the maintenance of the patient's health. Health advice is an example of support information. The "user" can include both medical staff and patients. There may be a plurality of client devices 11. The client device 11 may be a stationary device installed at a specific location, or a portable device that can be carried by a user.

[0017] The server device 12 is installed at a location remote from the client device 11. The client device 11 and the server device 12 are configured to perform two-way communication of data via a communication network 20.

[0018] As illustrated in FIG. 2, the client device 11 includes a communication interface 111. On the other hand, the server device 12 includes a communication interface 121. Each of the communication interface 111 and the communication interface 121 is a hardware interface configured to enable the above-described two-way data communication.

[0019] The client device 11 comprises a processor 112 and a user interface 113. The user interface 113 is a hardware interface that mediates the exchange of information between the user and the processor 112 and receives user instructions to cause the client device 11 to perform a predetermined operation. The processor 112 and the user interface 113 may be mounted in a common enclosure or provided in a distributed manner across separate enclosures.

[0020] The predetermined operation performed by the client device 11 includes the transmission of prescription information PR to the server device 12. Prescription information PR contains information related to the prescription of medication to the patient. Examples of information related to medication prescription include the following items. However, not all of the listed items need to be included in prescription information PR. • Date the medication was prescribed • Date the medication was dispensed based on the prescription • Date the medication history was finalized • Date the medication guidance was provided • Name of the drug, ingredients, therapeutic classification, dosage form, administration method, and number of days prescribed. • Identification information of medical institutions • Prescription identification information • Identification information of healthcare professionals • Patient identification information • Patient age, sex, and measurable biometric parameters • Patient's medical history, medication history, and usage of over-the-counter (OTC) medications and health foods.

[0021] The server device 12 comprises a processor 122 and an inference model 123. The inference model 123 is a computer algorithm executable by the processor 122. The inference model 123 is configured to take prescription information PR as input and output an index value related to the effectiveness of health advice to patients as an inference result.

[0022] The indicator values ​​may include items related to improving patients' medication adherence. Improved medication adherence is an example of behavioral change. These indicator values ​​are examples of indicator values ​​related to behavioral change. Since it is desirable for patients with chronic diseases to continue treatment, for each patient, indicator values ​​may include the total number of days prescribed medication within a specified period, the total number of prescriptions within a specified period, the total cost of medication within a specified period, the interval between visits to the medical institution, and the difference between the scheduled date of visit to the medical institution and the date on which the prescribed medication is completed.

[0023] Statistical values ​​of the above indicators for a specific group can also be examples of indicators of behavioral change. Examples of such groups include at least one healthcare institution, at least one healthcare worker, and at least one patient. Examples of statistical values ​​include the mean, median, mode, sum, maximum, and minimum.

[0024] Other examples of behavioral changes in patients as a result of health advice include undergoing tests, seeing a doctor, and consulting with a physician about changing medications.

[0025] Therefore, the rate of return visits to healthcare facilities for each patient can also be an example of an indicator value related to behavioral change. This indicator value may be obtained for a specific group. Examples of such groups include at least one patient who received health advice, at least one patient who did not receive health advice, at least one healthcare facility, and at least one healthcare professional. If background information such as age and medical history differs within a group of patients, groups may be formed based on the same or similar background information.

[0026] Indicators related to behavioral change may include items related to the management of healthcare institutions. Since the number of patients per day may increase as patient behavior changes in an improved direction, at least one of the numerical values ​​and statistical values ​​of the number of patients per given period, prescription amount, etc., can serve as indicator values. Examples of given periods include one day, a specific day of the week, at least one week, or at least one month.

[0027] As illustrated in Figure 1, the support system 10 includes an inference model generator 13. The inference models 123 of the server device 12 are generated through machine learning performed by the inference model generator 13. The inference model generator 13 is an example of a computing device.

[0028] As illustrated in Figure 3, the inference model generation device 13 is equipped with an input interface 131. The input interface 131 is configured as a hardware interface capable of accepting raw data OR used for machine learning from a database (not shown).

[0029] The database is a storage device that can be implemented using semiconductor memory, hard disk drives, magnetic tape drives, etc. This storage device may be installed in at least one of the client device 11 and the server device 12, or it may be installed in the inference model generation device 13. This storage device may also be a device independent of the client device 11 and the server device 12, capable of communicating with the inference model generation device 13 via the communication network 20.

[0030] The inference model generation device 13 includes a processor 132. The processor 132 is configured to extract predetermined features from advice information related to health advice previously provided to patients, which is included in the raw data OR received through the input interface 131, and to generate training data in which the features are associated with index values ​​related to at least one of the provision and effectiveness of the health advice corresponding to the advice information. The advice information is an example of support information.

[0031] The inference model generation device 13 is equipped with an output interface 133. The processor 132 is configured to generate an inference model 123 by performing machine learning to train a neural network with training data, and output it through the output interface 133. The inference model 123 output from the output interface 133 is implemented in the server device 12.

[0032] Figure 4 illustrates the process flow for generating training data, which is executed by the processor 132 of the inference model generation device 13.

[0033] It is preferable that the training data be generated to include time-series information. On the other hand, prescription issuance, dispensing, and provision of health advice are not carried out on holidays at medical institutions, and holidays vary depending on the medical institution. Processor 132 performs preprocessing by assigning a specific value (such as zero) to the data corresponding to holidays included in the original data OR (STEP 11).

[0034] Next, processor 132 extracts advice information contained in the original data OR (STEP 12) and obtains prescription information associated with the extracted advice information (STEP 13). Prescription information not associated with advice information may also be obtained as an example, including information such as "no health advice was given."

[0035] Next, the processor 132 extracts predetermined features from at least one of the advice information and the prescription information (STEP 14). Extraction may be performed over a predetermined period. Examples of predetermined periods include one day, a specific day of the week, at least one week, at least one month, etc.

[0036] Extraction may be performed not only for specific patients but also for specific groups. Examples of such groups include at least one healthcare institution, at least one healthcare professional, or patients with the same or similar background information. Examples of background information include gender, age, physical characteristics, disease, use of over-the-counter (OTC) medications and health supplements, type of prescribed medication (ingredients, therapeutic classification, dosage form, administration, etc.), and cumulative number of prescriptions.

[0037] In this example, relevant prescription information is identified based on the advice information. However, since multiple health advice may be associated with a particular drug or prescription, relevant advice information may also be identified based on the type of drug included in the prescription information.

[0038] The extracted features may include items necessary for considering seasonality. Examples of such items include the date and day of the week when the drug was prescribed, and the date and day of the week when the medication was dispensed based on the prescription.

[0039] The extracted features may include items related to the prescribed medication. Examples of such items include name, ingredients, therapeutic classification, dosage form, dose, administration method, and average prescription duration.

[0040] The extracted features may include items related to health advice. Examples of such items include identifying information that identifies the health advice, the type of target of the health advice (disease, treatment, etc.), the type of information included in the health advice (drug usage, side effects, severity of the disease, usefulness of the treatment, etc.), and the timing of the health advice (time of implementation, difference from the time used for inference, etc.).

[0041] The extracted features may include items related to the manner in which health advice is provided. Examples of such items include the method of providing health advice (in person, online, telephone, social media, messaging apps, postcards, etc.) and the amount of information included in the health advice (number of pieces of advice, number of characters, font size, presence or absence of diagrams or illustrations, etc.). Furthermore, if a message containing health advice is delivered via social media or an app, whether or not a link included in the message was selected may also be a feature.

[0042] The extracted features may include items related to the provision of health advice. Examples of such items include the number and rate of health advice provided, the number of prescriptions containing medications associated with the health advice, the prescription amount for medications associated with the health advice, the number and percentage of patients to whom health advice was provided, and the number and percentage of patients who visited a medical institution and had received health advice.

[0043] The extracted features may include items related to medical institutions. Examples of such items include identifying information that identifies a medical institution (corporation, medical department, store, size, etc.), geographical location information of a medical institution, identifying information that identifies medical personnel, the number of medical personnel involved in dispensing, prescription concentration rate, and indicators related to busyness and congestion (total number of prescriptions issued, number per medical personnel, time spent on medication guidance and its statistics, etc.).

[0044] The extracted features may include items related to the patient. Examples of such items include the cumulative number of visits to a specific medical institution, the interval between visits to the medical institution, the scheduled date of the next visit, the difference between the scheduled date and the actual date of the next visit and its statistical value, age, sex, physical characteristics, measured values ​​of biometric parameters, medical history, medication history, usage of over-the-counter (OTC) drugs and health foods, information on previously prescribed medications (name, ingredients, drug classification, dosage form, dose, usage, number of days prescribed, etc.), presence or absence of side effects, type of side effects, and information on health advice provided in the past (content, frequency, etc.).

[0045] Next, processor 132 calculates an index value related to the effectiveness of past health advice based on the extracted features and associates them (STEP 15). This generates training data in which specific features are associated with specific index values. There can be multiple combinations of features and index values. In addition, the calculated index values ​​and the time-series data of the index values ​​(changes in index values ​​over time) can also become features.

[0046] The behavioral changes observed in patients as a result of health advice can be verified in the following way.

[0047] Switching medications can be an example of behavioral change. Based on medication prescription data from the time health advice was received, it can be verified whether a switch to the medication recommended in that advice was made. Whether or not a switch was made can be used as an indicator of effectiveness.

[0048] The implementation of the test can be an example of behavioral change. Patients who receive health advice recommending the test can be given a questionnaire, and whether they actually underwent the test can be verified based on their responses. The questionnaire can be given at an appropriate time after the health advice is given. The questionnaire can be distributed as a paper form when medication is prescribed and collected when the patient returns to the medical institution. Responses can also be given to questionnaires distributed electronically via apps or social media. Whether or not the test was taken can be used as an indicator of effectiveness.

[0049] Consulting a doctor based on the next medication guidance plan (OP; Observation Plan) recorded in the patient's medical history as health advice can be an example of behavioral change. Whether the patient consulted a doctor based on the health advice can be verified through an interview conducted when the patient returns to the medical institution. Whether or not a consultation took place can be used as an indicator of effectiveness.

[0050] Next, processor 132 performs supervised learning using gradient boosting so that an inference model 123 is constructed that extracts the above features from the input patient prescription information PR and infers the above index values ​​(STEP 16). Other algorithms may be used for machine learning. Examples of other machine learning algorithms include neural networks, state-space models, linear regression, logistic regression, decision trees, random forests, and support vector machines.

[0051] Furthermore, new health advice for which no past data exists can also be included in the inference process. In this case, by training the model with trends from other health advice as training data, it may be possible to perform inference using disease type and prescription number as input.

[0052] Depending on the subject of the inference, supervised learning and time series analysis methods can be used interchangeably. If historical data is unavailable, supervised learning is used as described above; if historical data is available, time series analysis methods can be used. Both learning methods may also be used in combination. For example, inferences concerning a predetermined period can be performed using supervised learning, and then the obtained data can be used as input to perform inferences concerning the future using time series analysis methods.

[0053] As described above, the inference model 123 generated as described above is implemented in the server device 12. As illustrated in Figure 2, the processor 122 of the server device 12 is configured to input prescription information PR, which is sent from the client device 11 and received by the communication interface 121, into the inference model 123, and to send inference result information RS, which is associated with the index value output as an inference result from the inference model 123, to the client device 11 via the communication interface 121.

[0054] The processor 112 of the client device 11 is configured to provide the inference result information RS received via the communication interface 111 to the user interface 113. The user interface 113 can be implemented by a display device, a printing device, or the like.

[0055] Figure 5 illustrates a screen 30 displayed on a display device based on the inference result information RS. In this example, screen 30 includes multiple regions 31. Each region 31 is configured to display at least a portion of a health piece of advice, along with information corresponding to an index value related to the effectiveness of that health piece of advice. By selecting a region 31 in which a portion of the health piece of advice is displayed, the user can transition to a screen in which the entire health piece of advice can be displayed.

[0056] According to the configuration of this embodiment, the user of the client device 11 can obtain information regarding the effectiveness of health advice related to a prescription by inputting prescription information PR related to the prescription of a drug given to a patient, prior to providing the health advice. This can increase the motivation of healthcare professionals to provide health advice to patients.

[0057] Such health advice can encourage patients to become more proactive in their treatment, thereby promoting repeat visits to medical institutions. This can not only improve the patient's health but also improve the financial performance of the medical institution. Improved medication adherence among patients requiring treatment can also improve the financial performance of pharmaceutical companies.

[0058] For service providers offering health advice, obtaining information on the effectiveness of that advice allows them to identify highly useful advice. This, in turn, encourages the creation and provision of even higher-quality advice.

[0059] As illustrated in Figure 6, the processor 122 of the server device 12 may be configured to change the manner in which inference result information RS is provided according to the index values ​​output by the inference model 123. In this example, the area 31 that provides health advice with a higher implementation effect is displayed larger.

[0060] As an alternative, the system may be modified so that health advice with a relatively high effectiveness is displayed in a more eye-catching position on screen 30 (e.g., at the top or in the center).

[0061] As an alternative, multiple health advice pieces may be displayed on screen 30, arranged from top to bottom in order of their effectiveness.

[0062] Only health advice whose effectiveness exceeds a predetermined threshold may be displayed on screen 30. In other words, health advice with relatively low effectiveness may be hidden in the initial state. The user interface 113 of the client device 11 may be configured to allow the user to select whether to display or hide health advice according to the level of effectiveness.

[0063] In the case of medications used for various diseases, health advice related to various diseases may be associated with the same medication. In such cases, the health advice that is used most frequently may be displayed preferentially on screen 30.

[0064] If multiple medications are being used, health advice that is most frequently used based on the combination of medications may be displayed preferentially on screen 30.

[0065] The inference result information RS may include the methods and timing for providing highly effective health advice. If font size affects the effectiveness, the font size provided in the user interface 113 may be automatically changed.

[0066] If there are constraints on the duration and frequency of health advice sessions, or the number of people to whom health advice is provided via social media or messaging apps, the health advice that can maximize effectiveness within those constraints may be selected.

[0067] With the above configuration, health advice with high effectiveness can be prioritized, reducing the time spent selecting advice and streamlining medication guidance. For example, during peak hours at medical institutions, health advice can be prioritized for patients who are likely to benefit most from it, making efforts to improve medication adherence throughout the entire institution more efficient. For patients who are presumed to benefit more from non-face-to-face communication, actively using social media and messaging apps can help reduce congestion and waiting times at medical institutions. For patients, this leads to more opportunities to receive more beneficial health advice.

[0068] This study will reveal the relationship between medication prescriptions given to patients and effective health advice, thus offering educational benefits to healthcare professionals. By demonstrating effective methods and timing for providing health advice, it is also expected to improve healthcare professionals' communication skills with patients. Furthermore, it may assist in deciding whether to provide previously given health advice again.

[0069] In addition, it is expected to be useful as a reference for reviewing existing health advice and for considering the content of newly created health advice.

[0070] It is also possible to compare the effectiveness of health advice among multiple medical institutions belonging to the same group. If differences in effectiveness are observed, the causes and improvement measures can be investigated by analyzing the differences in features.

[0071] Each of the processors 112 of the client device 11, 122 of the server device 12, and 132 of the inference model generation device 13, which have the various functions described above, can be realized by at least one general-purpose microprocessor operating in cooperation with at least one general-purpose memory. Examples of general-purpose microprocessors include CPUs, MPUs, and GPUs. Examples of general-purpose memory include ROMs and RAMs. In this case, the ROM may store a computer program that performs the above-described processing. ROM is an example of a non-temporary computer-readable medium in which a computer program is stored. The general-purpose microprocessor selects at least a portion of the program stored in the ROM and loads it onto the RAM, and then works with the RAM to execute the above-described processing. The computer program may be pre-installed in the general-purpose memory, or it may be downloaded from an external server device via a communication network and then installed in the general-purpose memory. In this case, the external server device is an example of a non-temporary computer-readable medium in which a computer program is stored.

[0072] Each of processors 112, 122, and 132 may be implemented by at least one dedicated integrated circuit capable of executing the above-described computer program. Examples of dedicated integrated circuits include microcontrollers, ASICs, FPGAs, etc. In this case, the above-described computer program is pre-installed in a memory element included in the dedicated integrated circuit. The memory element is an example of a computer-readable medium in which the computer program is stored. Each of processors 112, 122, and 132 can also be implemented by a combination of a general-purpose microprocessor and a dedicated integrated circuit.

[0073] The configurations described herein are merely examples to facilitate understanding of this disclosure. Each configuration example may be modified or combined with other configuration examples as appropriate, as long as it does not deviate from the intent of this disclosure.

[0074] The inference models 123 generated by the inference model generation device 13 may be multiple. For example, an inference model 123 may be generated for each of several drugs with different prescription quantities. In this case, all generated inference models 123 are implemented in the server device 12, and the processor of the server device 12 may be configured to change the inference model 123 used for inference depending on the type of drug extracted from the prescription information PR.

[0075] Since the number of people prescribed a drug varies greatly depending on the drug, the scale of the number of people inferred as an indicator can also vary greatly. If an inference model with a prediction error of around 100 people is constructed, the impact of the prediction error on inference results based on drugs prescribed to tens of thousands of people cannot be ignored compared to inference results based on drugs prescribed to hundreds of people. With the above configuration, an inference model 123 capable of performing appropriate inference according to the type of drug extracted from prescription information PR is selected, thereby suppressing a decline in the quality of health advice provided based on the inference results.

[0076] In the above embodiment, an index value related to the effectiveness of health advice is inferred by the inference model 123. In addition to or instead of this, an index value related to the extent to which health advice is provided within a predetermined period may be inferred. Examples of a predetermined period include one day, a specific day of the week, at least one week, or at least one month.

[0077] Examples of such indicators include the number and rate of health advice provided, the number of prescriptions including medications associated with health advice, the amount of money spent on prescriptions for medications associated with health advice, the number and percentage of patients who receive health advice, and the number and percentage of patients who visit healthcare facilities who have previously received health advice.

[0078] In this case, the features used in the machine learning of the inference model 123 performed by the inference model generator 13 may include features related to education for healthcare professionals. The source data OR input to the inference model generator 13 may include information related to educational content for healthcare professionals. These features may include the usage status of the educational content (date of use, frequency of use, etc.).

[0079] With this configuration, when creating new health advice, it is possible to obtain in advance the results of an inference about how often that health advice will be implemented, which can support the decision of whether or not to create it. Furthermore, it can support the decision to prioritize the creation of health advice that is more likely to be used.

[0080] It is also possible to compare predictions of health advice implementation among multiple medical institutions belonging to the same group. If differences are found in the predicted implementation status, the causes can be investigated and improvement measures can be considered by analyzing the differences in features.

[0081] By training a large-scale language model with highly effective health advice, new health advice can be automatically generated by the model. Alternatively, recorded or videotaped conversations between healthcare professionals and patients can be converted into text information and used in conjunction with the training process. The automatically generated health advice can be reviewed and modified as needed by healthcare professionals or other relevant personnel.

[0082] In the examples described so far, the inference model 123 infers the index values ​​related to at least one of the implementation and effectiveness of health advice. However, the inference model 123 may also infer the index values ​​related to at least one of the implementation and effectiveness of medication guidance. Medication guidance is one example of support information provided to patients.

[0083] In this case, features similar to those related to the health advice mentioned above can be extracted from the original data OR regarding medication guidance and used in the machine learning of the inference model 123. The various features related to health advice may also be used in conjunction with the machine learning.

[0084] In the above embodiment, machine learning of the inference model 123 is performed by the processor 132 of the inference model generation device 13, which is independent of the client device 11 and the server device 12. However, machine learning of the inference model 123 may also be performed by the processor 112 of the client device 11 or the processor 122 of the server device 12. In this case, each of the processors 112 and 122 is an example of an arithmetic unit. [Explanation of symbols]

[0085] 11: Client device, 112: Processor, 12: Server device, 122: Processor, 123: Inference model, 13: Inference model generator, 132: Processor, PR: Prescription information, RS: Inference result information

Claims

1. A method for generating an inference model using a computing device, We prepared training data in which features extracted from support information provided to patients are associated with indicator values ​​related to at least one of the implementation and effectiveness of the provision of said support information. Machine learning is performed using the training data so that an inference model is constructed that extracts the features from prescription information related to drug prescriptions for input patients and infers the index value. Methods for generating inference models.

2. The inference model is generated for each of several drugs with different prescription volume scales. A method for generating an inference model according to claim 1.

3. A server device capable of communicating with client devices, An interface for receiving prescription information related to the prescription of medications to patients from the client device, A processor that infers an index value related to at least one of the implementation and effectiveness of providing support information to the patient based on the prescription information, and causes the client device to provide the inference result information associated with the index value, It is equipped with Server device.

4. The processor changes the manner in which the inference result information is provided according to the index value. The server device according to claim 3.

5. The processor modifies the inference model used to infer the index value according to the type of drug extracted from the prescription information. The server device according to claim 3 or 4.

6. A computer program that can be executed by a processor installed in a server device capable of communicating with a client device, By being executed, the server device will Prescription information related to the prescription of medication to patients is received from the client device. Based on the aforementioned prescription information, infer an indicator value related to at least one of the implementation and effectiveness of providing support information to the patient, The client device is provided with inference result information associated with the aforementioned index value. Computer program.

7. A method for supporting the provision of information to a patient, which is performed by a client device and a server device, Prescription information relating to the prescription of medication for the aforementioned patient is transmitted from the client device to the server device. In the server device, based on the prescription information, an index value relating to at least one of the implementation and effectiveness of providing support information to the patient is inferred, The server device transmits the inference result information associated with the aforementioned index value to the client device. In the aforementioned client device, the inference result information is provided by Methods for providing information.

Images