Computer program, information processing device, information processing method, and learning model generation method

Abstract

The invention provides a computer program, an information processing device, an information processing method, and a learning model generation method capable of assisting appropriate pain management. The computer program causes the computer to execute a process for acquiring pain-related information relating to the pain of the patient, inputting the acquired pain-related information to a learning model that outputs pain prediction information when the pain-related information has been input, acquiring the pain prediction information of the patient, and determining the pain prediction information of the patient on the basis of the acquired pain prediction information. And generating pain treatment assistance information for the patient, and outputting the generated pain treatment assistance information.

Description

Technical Field

[0001] The present invention relates to a computer program, an information processing device, an information processing method, and a learning model generating method. Background Art

[0002] Patent document 1 discloses the following diagnosis and treatment information management system, which has multiple treatment department servers that provide diagnosis and treatment information stored in the databases of multiple treatment departments within a medical institution, and a specific disease treatment unified server that obtains specific disease diagnosis and treatment information from the treatment department servers and provides unified specific disease diagnosis and treatment information from a unified database. By being able to refer to the diagnosis and treatment information of specific diseases such as cancer in all therapies that combine surgery, radiotherapy, chemotherapy and palliative care, and registered related information on a wide range of specific diseases, the progress of treatment can be grasped in a highly real-time manner.

[0003] Patent Document 1: Japanese Patent Application Publication No. 2010-3135

[0004] There are short-term prognostic prediction methods for diseases such as cancer (such as PPI: Palliative Prognostic Index), but this method is about the prognosis of the symptoms and does not specifically target the pain caused by the disease. In addition, for doctors, there is also the problem that they can only evaluate pain when they are in front of the patient. In addition, effective pain management requires specialized knowledge, and for doctors other than specialists, there is a problem that it is difficult to provide accurate pain treatment according to the patient. Summary of the invention

[0005] The present invention has been made in view of such circumstances, and an object of the present invention is to provide a computer program, an information processing device, an information processing method, and a learning model generating method that can assist appropriate pain treatment.

[0006] (1) The computer program of the present invention enables a computer to perform the following processing: obtaining pain-related information related to a patient's pain, inputting the obtained pain-related information into a learning model that outputs prediction information of pain when pain-related information is input, so as to obtain prediction information of the patient's pain, generating pain treatment auxiliary information for the patient based on the obtained pain prediction information, and outputting the generated pain treatment auxiliary information.

[0007] Here, for the embodiments of the present invention,

[0008] (2) In the computer program of (1) above, preferably, the pain-related information includes at least one of an evaluation index for evaluating pain, vital data, medication information, and physical condition.

[0009] (3) The computer program of (1) or (2) above preferably causes the computer to perform the following processing: obtaining medical information related to the diagnosis and treatment of the patient, and then inputting the obtained medical information into the learning model that outputs prediction information of pain when the medical information is input, so as to obtain prediction information of the patient's pain.

[0010] (4) In any one of the computer programs (1) to (3) above, the medical information preferably includes at least one of diagnosis information, examination information, treatment information, prescription information, and disease prognosis prediction information.

[0011] (5) Any one of the computer programs (1) to (4) above preferably causes a computer to perform the following processing: obtaining literature information related to pain treatment, and then inputting the obtained literature information into the learning model that outputs prediction information of pain when the literature information related to pain treatment is input, so as to obtain prediction information of the patient's pain.

[0012] (6) In any one of the computer programs (1) to (5) above, the pain treatment support information preferably includes at least one of recommended prescription information and recommended treatment information for supporting a doctor examining the patient.

[0013] (7) In any one of the computer programs described in (1) to (6) above, the pain treatment support information preferably includes a method for supporting the patient's coping skills.

[0014] (8) Any one of the computer programs (1) to (7) above preferably causes a computer to execute a process of outputting a pain prediction pattern indicating a temporal transition of a pain degree based on prediction information of pain of the patient.

[0015] (9) The information processing device of the present invention comprises: a first acquisition unit, which acquires pain-related information related to the patient's pain; a second acquisition unit, which inputs the acquired pain-related information into a learning model that outputs prediction information of pain when pain-related information is input, so as to acquire prediction information of the patient's pain; a generation unit, which generates pain treatment auxiliary information for the patient based on the acquired pain prediction information; and an output unit, which outputs the generated pain treatment auxiliary information.

[0016] (10) The information processing method of the present invention obtains pain-related information related to the patient's pain, inputs the obtained pain-related information into a learning model that outputs prediction information of pain when pain-related information is input, so as to obtain prediction information of the patient's pain, generates pain treatment auxiliary information for the patient based on the obtained pain prediction information, and outputs the generated pain treatment auxiliary information.

[0017] (11) The learning model generation method of the present invention obtains first training data, which includes: pain-related information related to pain of multiple patients and predicted information on the pain of the multiple patients, and generates a learning model based on the first training data obtained in a manner that outputs predicted information on pain when pain-related information is input.

[0018] (12) The learning model generation method of (11) above is preferably a method in which second training data is obtained in the learning model generation method of (11) above, and the second training data also includes: diagnosis and treatment information of multiple patients, literature information related to pain treatment, and prediction information of pain for the multiple patients. Based on the obtained second training data, the learning model is generated in a manner that outputs prediction information of pain when the diagnosis and treatment information and literature information related to pain treatment are input.

[0019] According to the present invention, appropriate pain treatment can be assisted. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a diagram showing an example of the configuration of the pain treatment support system according to the present embodiment.

[0021] Figure 2 This is a diagram showing an example of the structure of a pain treatment support device.

[0022] Figure 3 is a diagram showing information collection by means of a pain treatment assisting device.

[0023] Figure 4A This is a diagram showing an example of information in a database.

[0024] Figure 4B This is a diagram showing an example of information in a database.

[0025] Figure 4C This is a diagram showing an example of information in a database.

[0026] Figure 5 This is a diagram showing provision of pain treatment support information by a pain treatment support device.

[0027] Figure 6 1 is a diagram showing a first example of a process for generating pain treatment support information by a pain treatment support device.

[0028] Figure 7 1 is a diagram showing a second example of the process of generating pain treatment support information by the pain treatment support device.

[0029] Figure 81 is a diagram showing a third example of the process of generating pain treatment support information by the pain treatment support device.

[0030] Fig. 9 This is a diagram showing an example of a method for generating a learning model.

[0031] Fig.10 This is a diagram showing a first example of pain treatment support information on the doctor side.

[0032] Fig.11 This is a diagram showing a second example of pain treatment support information on the doctor side.

[0033] Fig.12 This is a diagram showing a third example of pain treatment support information on the doctor side.

[0034] Fig.13 This is a diagram showing a fourth example of pain treatment support information on the doctor side.

[0035] Fig.14 This is a diagram showing a first example of pain treatment support information on the patient side.

[0036] Fig.15 This is a diagram showing a second example of pain treatment support information on the patient side.

[0037] Fig.16 This is a diagram showing an example of a processing sequence using a pain treatment support device.

[0038] Fig.17 This is a diagram showing an example of the generation order of the learning model. DETAILED DESCRIPTION

[0039] Hereinafter, embodiments of the present invention will be described. Figure 1 : is a diagram showing an example of the structure of the pain treatment assistance system of the present embodiment. The pain treatment assistance system as an information processing system includes a pain treatment assistance device 50 as an information processing device. The pain treatment assistance system may also include a patient terminal device 10, a device 20, an assistant terminal device 30, and a doctor terminal device 40. The patient terminal device 10, the assistant terminal device 30, and the doctor terminal device 40 are connected to the pain treatment assistance device 50 via a communication network 1. The device 20 may be configured to be connected to the patient terminal device 10 wirelessly or by wire, or may be configured to be connected to the pain treatment assistance device 50 via the communication network 1. In addition, the device 20 is not essential, and for example, it may be configured to be composed only of input from the patient terminal device 10.

[0040] The patient terminal device 10 is a terminal device carried or possessed by the patient, and the assistant terminal device 30 is a terminal device carried or possessed by assistants such as the patient's family members and caregivers. The caregivers may be family members or caregivers in a nursing office. The patient terminal device 10 and the assistant terminal device 30 can be composed of a smartphone, a tablet computer, a personal computer, etc. having a display panel, an operation panel, a microphone, a speaker, etc. An application (application) for using the pain treatment assistance system is introduced into the patient terminal device 10 and the assistant terminal device 30. For the patient, for example, it is assumed that the patient suffers from cancer, etc., and regularly goes to a medical institution while recuperating at home, but it is not limited to such patients. In addition, the application can be a Web application or a download application.

[0041] Device 20 is a measuring device for measuring the patient's vital data, and can be a wearable terminal or a fixed device. Device 20 can measure vital data such as body temperature, heartbeat, brain wave, EDA (electrodermal activity: galvanic skin reflex). In addition, vital data is not limited to these, and any data that is considered to be related to the patient's pain caused by the disease can also be included. For example, vital data can also include blood pressure, pulse pressure, pulse, blood sugar level, weight, number of steps, activity level, etc.

[0042] The doctor terminal device 40 is a terminal device used by doctors in medical institutions, etc. The doctor terminal device 40 can be composed of a personal computer with a display panel, an operation panel, etc. A healthcare worker, a nurse, or other medical practitioner can also use the doctor terminal device 40 under the guidance of a doctor. An application (such as a Web application, etc.) for using the pain treatment support system is introduced into the doctor terminal device 40.

[0043] Figure 2 : is a diagram showing an example of the structure of the pain treatment auxiliary device 50. The pain treatment auxiliary device 50 includes a control unit 51 for controlling the entire device, a communication unit 52, a memory 53, a pain-related information DB 54, a diagnosis and treatment information DB 55, a pain treatment auxiliary information DB 56, a literature information DB 57, an auxiliary information generation unit (generation unit) 58, and a storage unit 59. The pain treatment auxiliary device 50 may also disperse the processing functions and be composed of a plurality of devices. Databases such as the pain-related information DB 54, the diagnosis and treatment information DB 55, the pain treatment auxiliary information DB 56, and the literature information DB 57 may be configured to be built into the pain treatment auxiliary device 50, or may be installed in an external data server. In addition, the storage unit 59 may also be installed in an external data server.

[0044] The storage unit 59 can be composed of, for example, a hard disk or a semiconductor memory, and stores a computer program 60 (program product), a learning model 61, and necessary information. The computer program 60 can be downloaded from an external device via the communication unit 52 and stored in the storage unit 59. In addition, the computer program 60 recorded in a recording medium (for example, an optically readable disk storage medium such as a CD-ROM) can be read by a recording medium reading unit and stored in the storage unit 59, or the computer program 60 can be read by a recording medium reading unit and expanded in the memory 53.

[0045] The control unit 51 is configured to be equipped with a required number of CPUs (Central Processing Units), MPUs (Micro-Processing Units), GPUs (Graphics Processing Units), etc. The control unit 51 is capable of executing the processing specified by the computer program 60. That is, the processing performed by the control unit 51 is also the processing performed by the computer program 60. The control unit 51 is capable of executing the functions of the auxiliary information generation unit 58 by executing the computer program 60. The auxiliary information generation unit 58 can be configured by hardware, can be configured to be implemented by software, or can be implemented by a combination of hardware and software. The control unit 51 is capable of performing processing using the learning model 61. In addition, the control unit 51 has the functions of a first acquisition unit, a second acquisition unit, a generation unit, and an output unit.

[0046] The communication unit 52 includes a communication module and can exchange necessary information with the patient terminal device 10, the assistant terminal device 30, and the doctor terminal device 40 via the communication network 1. The communication unit 52 can also exchange necessary information with the device 20.

[0047] The memory 53 can be formed of a semiconductor memory such as SRAM (Static Random Access Memory), DRAM (Dynamic Random Access Memory), or flash memory. The control unit 51 can expand the computer program 60 in the memory 53 and execute the computer program 60 .

[0048] Figure 3: is a diagram showing information collection by the pain treatment supporting device 50. The pain treatment supporting device 50 obtains pain-related information related to the patient's pain caused by the disease from the patient terminal device 10, the equipment 20, and the assistant terminal device 30 via the communication unit 52. The pain-related information is appropriately repeatedly sent from the patient terminal device 10, the equipment 20, and the assistant terminal device 30. For example, the pain-related information may be sent once during the period from the patient's last visit to the hospital to the next visit to the hospital, or may be sent multiple times depending on the length of the period. The pain-related information obtained is classified for each patient and at each acquisition time and stored in the pain-related information DB 54.

[0049] The pain treatment support device 50 also acquires medical information related to the patient's medical treatment from the doctor terminal device 40 via the communication unit 52. The acquired medical information is classified for each patient and for each medical treatment time and stored in the medical information DB 55.

[0050] Figure 4A , 4B And 4C is a diagram showing an example of information in the database. Figure 4A An example of information in the pain-related information DB 54 is shown. Pain-related information can be classified into pain evaluation index (evaluation index for evaluating pain), vital data, medication information, and physical condition. The pain-related information sent from the patient terminal device 10 and the device 20 includes at least one of the pain evaluation index, vital data, medication information, and physical condition. In addition, the classification is an example and is not limited to Figure 4A Example.

[0051] The pain evaluation index includes the intensity of pain, changes in pain, the timing of pain, etc. The pain evaluation index can be sent from the patient terminal device 10 to the pain treatment support device 50, for example, by the patient starting an application of the patient terminal device 10 and answering questions related to the intensity of pain, changes in pain, the timing of pain, etc.

[0052] The vital data include data that changes due to the response to pain, such as body temperature, heart rate, brain wave, EDA, pulse, etc. In addition, it may also include blood pressure, pulse pressure, blood sugar level, weight, number of steps, amount of activity, etc. The vital data can be acquired using the device 20.

[0053] The medication information includes the type of medication, the amount of medication, the medication period, the medication time (timing), whether medication is taken, etc. The medication information may include medications that have been taken in the past, not just medications that are currently being taken.

[0054] Physical conditions include activity level, sleep time, sleepiness, fatigue, delusions, constipation, voice, and expression, etc. It can include items that can evaluate whether there is a disorder due to pain, medicine, disease symptoms, etc.

[0055] The pain-related information includes at least one of a pain evaluation index (an evaluation index for evaluating pain), vital data, medication information, and physical condition.

[0056] By acquiring the pain-related information as described above from the patient terminal device 10 and the equipment 20 , the patient can accurately communicate the pain and symptoms he or she feels at a necessary timing and frequency during the period from the previous examination to the next examination.

[0057] Figure 4B An example of information in the medical information DB 55 is shown. Medical information can be classified into diagnosis information, examination information, prescription information, treatment information, and disease prognosis prediction information. The medical information sent from the doctor terminal device 40 includes at least one of diagnosis information, examination information, prescription information, treatment information, and disease prognosis prediction information. In addition, the classification is an example and is not limited to Figure 4B Example.

[0058] Diagnostic information includes the results of observations during the medical examination, such as the interview, visual examination, palpation or percussion, and electronic medical records. Examination information includes blood tests, urine tests, and imaging examinations. Prescription information includes the type and amount of medicine prescribed by the doctor. In addition, it may also include the period during which the medicine is taken. Treatment information includes treatment methods and the period of treatment. In addition, pain treatment includes not only literal treatment but also prescriptions for medicines. Disease prognosis prediction information includes PPI (Palliative Prognostic Index). PPI is a representative indicator for predicting short-term life prognosis (weekly units), and uses a score that is the sum of the scores of PPS (Palliative Performance Scale), oral intake, edema, dyspnea at rest, and delusions. The larger the score, the shorter the predicted prognosis.

[0059] The diagnosis and treatment information includes at least one of diagnosis information, examination information, treatment information, prescription information and disease prognosis prediction information.

[0060] Figure 4C An example of information in the document information DB57 is shown. Document information can be classified into guidelines and papers. Guidelines are guidelines related to pain treatment, for example, including guidelines indicating what kind of pain treatment is recommended for the patient's symptoms and pain. Papers are papers related to pain treatment, for example, including objective information indicating what kind of pain treatment is effective for the patient's symptoms and pain. Document information DB57 is updated appropriately to record the latest information.

[0061] Figure 51 is a diagram showing the provision of pain treatment assistance information by means of a pain treatment assistance device 50. The pain treatment assistance device 50 (control unit 51) transmits the pain treatment assistance information generated by the processing described later to the patient terminal device 10, the assistant terminal device 30, and the doctor terminal device 40. The transmission can be performed at a necessary time and with a necessary frequency. In addition, information that is preferably provided to the patient's assistant among the pain treatment assistance information provided to the patient side can also be transmitted to the assistant terminal device 30. In addition, information that is preferably not provided to the patient but only to the family members among the pain treatment assistance information provided to the patient side can also be transmitted to the assistant terminal device 30 instead of being transmitted to the patient terminal device 10.

[0062] Figure 6 1 is a diagram showing a first example of the generation process of pain treatment auxiliary information by means of a pain treatment auxiliary device 50. The control unit 51 reads pain-related information of a specific patient from the pain-related information DB54, and inputs the read pain-related information into a learning model 61. The pain-related information of a specific patient is time series data of one or more moments. For example, it may include not only the current pain-related information of the specific patient, but also pain-related information of past moments traced back from the present. The control unit 51 is capable of performing a vectorization operation on the patient's pain-related information to generate a unique vector to generate a pain-related information vector, and inputs the generated pain-related information vector into the learning model 61. In addition, a pain-related information vector may be generated for each of a plurality of moments. In addition, although not shown in the figure, the past pain pattern of the pain treatment auxiliary information DB56 may also be input into the learning model 61.

[0063] The learning model 61 is generated (learned) to output prediction information of pain (pain prediction information) when pain-related information is input. Pain-related information is time series data, for example, pain-related information within a period of one month, three months, six months or 12 months from the present retrospectively. The prediction information of pain is information representing the pain within a necessary period (for example, one month, three months, six months, 12 months, etc.) from the present moment to the future. For example, the learning model 61 can output pain prediction information from the present to one month later based on the pain-related information of the past three months, or it can output pain prediction information from the present to three months later based on the pain-related information of the past six months. In addition, the above-mentioned period is an example and is not limited to these. The intensity (degree) of pain can be expressed in a quantitative level, for example, in 10 levels such as 1 to 10 or in 5 levels such as 1 to 5. It can be set that the larger the numerical value, the stronger the patient's pain.

[0064] The learning model 61 can use, for example, SVM (Support Vector Machine), decision tree, random forest, adaptive boosting, neural network (for example, RNN: Recurrent Neural Network, LSTM: Long Short Term Memory, Transformer, etc.), etc.

[0065] The control unit 51 reads the patient's past pain pattern recorded in the pain treatment auxiliary information DB 56. The pain pattern is information about past pain, and is information indicating the transition of pain intensity, similar to the pain prediction information. The difference between the pain prediction information and the past pain pattern is that the former is the transition of pain intensity from the present to the future, while the latter is the transition of pain intensity from the present to the past. The control unit 51 inputs the read past pain pattern and the pain prediction information obtained from the learning model 61 into the auxiliary information generation unit 58.

[0066] The auxiliary information generation unit 58 generates pain treatment auxiliary information based on the input pain prediction information and past pain patterns. The auxiliary information generation unit 58 stores the generated pain treatment auxiliary information in the pain treatment auxiliary information DB 56. The pain treatment auxiliary information includes, for example, the pain prediction pattern, recommended prescription information and recommended treatment information for assisting doctors who examine patients, and recommended treatment information for assisting patients. The pain treatment auxiliary information only needs to include at least one of the recommended prescription information and recommended treatment information for assisting doctors.

[0067] The prediction pattern of pain is generated based on the prediction information of pain, for example, any information indicating the time transition of the patient's pain level from the past to the future. The prediction pattern of pain can be displayed by appropriate methods such as graphical display and table display. Recommended prescription information is information for assisting doctors, including information for determining the recommended prescription in the prescription of medicine for the patient. Recommended treatment information is information for assisting doctors, including information for determining the recommended treatment method in the treatment method other than the prescription of medicine for the patient. Recommended response information is information for assisting patients, for example, information for realizing that in the event of expected pain in the future, by notifying the patient or at least one of his family members and caregivers in advance, the patient can respond even without the intervention of a doctor such as a doctor's examination.

[0068] The auxiliary information generating unit 58 may generate the pain treatment auxiliary information in the rule base using, for example, a table showing the correspondence between the pain prediction information and the past pain pattern and the pain treatment auxiliary information. The table may also show the correspondence between the pain prediction information and the past pain pattern and the pain treatment auxiliary information.

[0069] The auxiliary information generating unit 58 can be configured by a model using an algorithm obtained by performing machine learning, and for example, SVM (Support Vector Machine), decision tree, random forest, adaptive boosting, neural network, etc. can be used.

[0070] As described above, the control unit 51 is able to obtain pain-related information related to the patient's pain, input the obtained pain-related information into the learning model 61 that outputs prediction information of pain when pain-related information is input to obtain prediction information of the patient's pain, generate pain treatment auxiliary information for the patient based on the obtained prediction information of pain, and output the generated pain treatment auxiliary information.

[0071] Figure 7 2 is a diagram showing a second example of the process of generating pain treatment support information by the pain treatment support device 50. Figure 6 The difference of the first example illustrated in the example is that in addition to the pain-related information, the medical information and the literature information are also input to the learning model 61. The control unit 51 reads the medical information of a specific patient from the medical information DB55, and inputs the read medical information into the learning model 61. The medical information of a specific patient is time series data of one or more moments. For example, it can also include not only the current medical information of a specific patient, but also the medical information of the past moments traced back from the present. The control unit 51 can generate a medical information vector by performing a vectorization operation to generate a unique vector on the medical information, and input the generated medical information vector into the learning model 61. In addition, the medical information vector can also be generated for each of multiple moments. For example, the learning model 61 can also output the pain prediction information from the current to one month later based on the pain-related information and medical information of the past three months, and can also output the pain prediction information from the current to three months later based on the pain-related information and medical information of the past six months. In addition, the above period is an example and is not limited to these. In addition, although not shown in the figure, the past pain pattern of the pain treatment auxiliary information DB56 can also be input into the learning model 61.

[0072] In addition, the control unit 51 reads the document information from the document information DB 57, and inputs the read document information into the learning model 61. The control unit 51 can generate a document information vector by performing a vectorization operation to generate a unique vector on the document information, and input the generated document information vector into the learning model 61. The document information, for example, only needs to include information required to establish an association between the patient's future pain condition and the changes that have occurred under the historical information such as the patient's symptom history, vital data history, examination history, medication history, and treatment history. For example, if there is a paper indicating how the pain will change in the future if a certain specific medicine is taken, the paper can be used as the document information. In addition, if there is a guide indicating how the pain will develop in the future if a certain specific treatment is implemented, the guide can be used as the document information. In particular, in the case where the patient's pain-related information and diagnosis and treatment information are insufficient (for example, in the case of less past data), the learning model 61 can output the patient's pain prediction information by using the document information to supplement the shortage of pain-related information and diagnosis and treatment information. By increasing the types of information input to the learning model 61, it can be expected that the accuracy of the index output by the learning model 61 will be improved.

[0073] In addition, the control unit 51 reads the document information from the document information DB 57, and inputs the read document information into the learning model 61. The control unit 51 can perform a vectorization operation to generate a unique vector on the document information to generate a document information vector, and input the generated document information vector into the learning model 61. By increasing the types of information input into the learning model 61, it can be expected that the accuracy of the index output by the learning model 61 will be improved.

[0074] As described above, the control unit 51 may acquire medical treatment information related to the patient's medical treatment and input the acquired medical treatment information into the learning model 61 that outputs prediction information of pain when the medical treatment information is input, thereby acquiring prediction information of pain of the patient.

[0075] Alternatively, the control unit 51 may acquire literature information related to pain treatment and input the acquired literature information to the learning model 61 that outputs prediction information of pain when the literature information related to pain treatment is input, thereby acquiring prediction information of pain of the patient.

[0076] Figure 8 2 is a diagram showing a third example of the generation process of pain treatment support information by the pain treatment support device 50. Figure 7 The second example illustrated in FIG. 5 is different in that the patient's pain-related information, diagnosis and treatment information, and literature information are input to the auxiliary information generating unit 58 in addition to the pain prediction information and past pain patterns. In addition, although not shown, past pain patterns of the pain treatment auxiliary information DB 56 may also be input to the learning model 61.

[0077] The auxiliary information generating unit 58 may generate the pain treatment auxiliary information in the rule base, for example, using a table indicating the correspondence between the pain prediction information, the past pain pattern, the patient's pain-related information, the diagnosis and treatment information and the literature information, and the pain treatment auxiliary information. The table may also indicate the correspondence between the pain prediction information, the past pain pattern, the patient's pain-related information, the diagnosis and treatment information and the literature information, and the pain treatment auxiliary information.

[0078] In addition, similar to the first and second examples, the auxiliary information generating unit 58 can be configured by a model using an algorithm obtained by performing machine learning, for example, SVM (support vector machine), decision tree, random forest, adaptive boosting, neural network, etc. can be used.

[0079] Fig. 9: is a diagram showing an example of a method for generating a learning model 61. In the following, it is assumed that the pain treatment supporting device 50 generates the learning model 61, but it is also possible to assume that a learning processing device other than the pain treatment supporting device 50 generates the learning model 61, and the pain treatment supporting device 50 obtains the generated learning model 61. The control unit 51 obtains training data (second training data) including pain-related information, medical information, literature information related to pain treatment, and prediction information of pain of multiple patients. These training data include disease course information and pain change information of multiple patients. The pain-related information and medical information are time series data, for example, including pain-related information within the past one month, three months, six months, or 12 months from the present. The literature information, for example, only needs to include information necessary to establish a correlation between the patient's future pain condition and the historical information such as the patient's symptom history, vital data history, examination history, drug taking history, and treatment history. For example, if there is a paper indicating how the pain will change in the future if a certain specific drug is taken or a certain specific treatment is performed, the paper can be used as the literature information. In addition, if there is guidance indicating how pain will change in the future if a certain specific treatment is implemented, the guidance can be used as literature information. The control unit 51 can generate a learning model 61 based on the acquired training data so that when pain-related information, diagnosis and treatment information, and literature information related to pain treatment are input, pain prediction information is output. The learning model 61 can be generated, for example, to output pain prediction information from the current to one month later based on pain-related information and diagnosis and treatment information from the past three months, or it can be generated to output pain prediction information from the current to three months later based on pain-related information and diagnosis and treatment information from the past six months. When generating the learning model 61, the parameters of the learning model 61 can be adjusted in such a way that the pain prediction information output by the learning model 61 is close to the teacher data included in the training data, that is, the pain prediction information.

[0080] In addition, Fig. 9 In the example of , the input data for learning includes pain-related information, diagnosis and treatment information, and literature information, but for example, the input data for learning may not include at least one of the diagnosis and treatment information and the literature information. For example, the control unit 51 may also obtain training data (first training data) including pain-related information related to pain of multiple patients and prediction information of pain of the multiple patients, and based on the obtained training data, generate a learning model 61 so that when pain-related information is input, prediction information of pain is output. However, if at least one of the diagnosis and treatment information and the literature information is included in the input data for learning, it is more likely that the inference accuracy of the learning model 61 will be improved.

[0081] Next, a specific example of pain treatment auxiliary information is described. The control unit 51 can output pain treatment auxiliary information to the doctor terminal device 40, the patient terminal device 10, or the assistant terminal device 30. The pain treatment auxiliary information is displayed on the display panel of the doctor terminal device 40, and the doctor can perform input operations and selection operations based on the displayed content. Similarly, the pain treatment auxiliary information is displayed on the display panel of the patient terminal device 10 or the assistant terminal device 30, and the patient, his family members, nursing staff, etc. can perform input operations and selection operations based on the displayed content.

[0082] Fig.10 400 is a diagram showing a first example of pain treatment support information 400 on the doctor's side. The pain treatment support information 400 displays patient information (patient ID, patient name, etc.), medication status (type of medication, medication amount, medication period, etc.), treatment status (treatment method, treatment period, etc.), and pain prediction pattern. Fig.10 In the example, the predicted pain pattern is displayed as a line graph, showing the progression of the patient's pain from the past to the future. The doctor can read that the pain is temporarily relieved at the current moment in the patient's pain treatment so far, but the pain is expected to gradually increase in the future.

[0083] In addition, as advice (supplementary information) to the doctor, “Pain may occur in the future.” When the doctor operates the “recommended prescription information” icon 401 , recommended prescription information described later can be displayed.

[0084] As described above, the control unit 51 outputs the predicted pattern of the patient's pain to the doctor terminal device 40 .

[0085] Fig.11 4 is a diagram showing a second example of pain treatment support information 410 on the doctor side. The pain treatment support information 410 includes recommended prescription information. The recommended prescription information displays patient information and recommended contents. Fig.11 In the example, as recommendation 1, "Take ×× pills of medicine ○○" is displayed, recommending taking a specific medicine. As recommendation 2, "Change medicine ○○ to △△" is displayed, recommending a change of medicine. In addition to the recommended content, "Recommend any one of the following measures" is also displayed, reminding people to take the prescription recommended by the doctor. Fig.11 The recommended prescription information shown in the figure can store a table corresponding to a plurality of pain prediction patterns in the storage unit 59 in advance. For example, the recommended prescription information R1, R2, R3, ... corresponding to the pain prediction patterns P1, P2, P3, ... can be selected from the table by the control unit 51. Fig.10 The pain prediction pattern closest to the pain prediction pattern of the patient shown is determined to establish recommended prescription information corresponding to the selected pain prediction pattern.

[0086] As described above, based on the patient's predicted pain pattern, the frequency and period of future pain can be predicted, and as a doctor, suitable analgesics can be suggested to the patient in advance. By taking the prescribed analgesics, the patient's quality of life (QOL) can be improved when the pain predicted to occur in the future does not occur. In this way, appropriate pain treatment can be assisted.

[0087] Fig.12 4 is a diagram showing a third example of the doctor's pain treatment support information 420. Fig.10 Compared with the first example, in the third example, the state where the pain value remains high has been passed from the past to the present. The doctor can read that the pain has not been relieved in the patient's pain treatment so far.

[0088] In addition, as advice (supplementary information) to the doctor, “The pain seems to continue in the future. How about studying new treatment methods?” If the doctor operates the “recommended treatment information” icon 421 , recommended treatment information described later can be displayed.

[0089] Fig.13 4 is a diagram showing a fourth example of pain treatment support information 430 on the doctor side. The pain treatment support information 430 includes a recommended treatment method. The recommended treatment method displays patient information and recommended contents. Fig.13 In the example, as recommendation 1, treatment method AA is displayed, and as recommendation 2, treatment method BB is displayed. In addition to the recommended content, it is also displayed as "Introducing experienced doctors in the absence of actual results of implementing the recommended treatment methods." If the doctor operates the "display doctor list" icon 431, the introduction information of doctors who have experience in recommending treatment methods will be displayed in the form of a list. The introduction information can include information such as the name of the medical institution, the name of the doctor, the treatment methods that the doctor is experienced in, and contact information. In addition, it can also be set to be able to contact the required doctor by operating the "contact" icon. For example Fig.13 The recommended treatment method shown in the figure can store a table corresponding to a plurality of pain prediction patterns in advance in the storage unit 59. For example, the recommended treatment methods T1, T2, T3, ... are corresponding to the pain prediction patterns P1, P2, P3, ... The control unit 51 can select the recommended treatment method T1, T2, T3, ... from the table. Fig.12 The pain prediction pattern that is closest to the pain prediction pattern of the patient shown is determined, and a recommended treatment method corresponding to the selected pain prediction pattern is determined.

[0090] Fig.14 1 is a diagram showing a first example of pain treatment support information 100 on the patient side. The pain treatment support information 100 shows a predicted pattern of pain. Fig.14In the example, the predicted pain pattern is displayed as a line graph, showing the patient's pain transition from the past to the future. The patient can read that the pain is temporarily relieved at the current moment, but the pain is expected to gradually increase in the future.

[0091] In addition, as advice (support information) to the patient or the assistant, “Pain is expected. Please confirm the countermeasure.” If the patient or the assistant operates the “Confirm” icon 101 , the countermeasure described later can be displayed.

[0092] As described above, the control unit 51 outputs the predicted pattern of the patient's pain to the patient terminal device 10 and the assistant terminal device 30 .

[0093] Fig.15 1 is a diagram showing a second example of the pain treatment support information 110 on the patient side. The pain treatment support information 110 includes a countermeasure (recommended countermeasure information). Fig.15 In the example, it is displayed as "Please take the following medicine. Take ×× pills of ○○. Please take it three times a day after each meal." Fig.15 The recommended treatment information shown in the figure can store a table corresponding to a plurality of pain prediction patterns in the storage unit 59 in advance. For example, the recommended prescription information S1, S2, S3, ... corresponding to the pain prediction patterns P1, P2, P3, ... can be selected from the table by the control unit 51. Fig.14 The pain prediction pattern closest to the patient's pain prediction pattern shown in the figure determines the recommended response information corresponding to the selected pain prediction pattern. The doctor can also predetermine the recommended prescription information S1, S2, S3, ... corresponding to the pain prediction patterns P1, P2, P3, ...

[0094] As described above, based on the predicted pattern of the patient's pain, the patient can know in advance how the frequency and period of pain will occur in the future. Then, by taking necessary coping methods based on the coping methods recommended by the pain treatment auxiliary device 50, the occurrence of pain can be suppressed in advance, and the patient's quality of life (QOL) can be improved. In particular, the patient can take necessary coping methods without going to the hospital, for example, at home. In this way, appropriate pain treatment can be assisted. In addition, in a case where it is not preferred to directly notify the patient of the pain treatment auxiliary information on the patient side, the pain treatment auxiliary information can also be sent to the assistant terminal device 30 instead of sending it to the patient terminal device 10.

[0095] As described above, according to this embodiment, the pain that the patient may experience in the future can be predicted based on the patient's condition, and pain treatment for alleviating the pain can be performed on the patient in advance. By taking preventive measures in this way, the patient's pain can be alleviated.

[0096] Fig.16 5 is a diagram showing an example of a processing sequence by means of the pain treatment auxiliary device 50. In the following, for convenience, the main body of the processing is described as the control unit 51. The control unit 51 obtains the patient's pain-related information (S11), and stores the obtained pain-related information in the pain-related information DB54 (database) (S12). The pain-related information can be repeatedly obtained for each patient at a necessary time. For example, it can be obtained once or multiple times during the period from the last visit to the hospital to the next visit to the hospital.

[0097] The control unit 51 acquires the patient's medical information (S13), and stores the acquired medical information in the medical information DB55 (database) (S14). The medical information can be acquired when the doctor determines that it is necessary, not only during the examination.

[0098] The control unit 51 inputs the patient's pain-related information, medical information, and literature information related to pain treatment into the learning model 61, and obtains the patient's pain prediction information output by the learning model 61 (S15). The control unit 51 reads the patient's past pain pattern from the pain treatment auxiliary information DB 56 (database) (S16).

[0099] The control unit 51 generates pain treatment auxiliary information (S17) including the predicted pattern of the patient's pain based on the patient's past pain pattern read from the pain treatment auxiliary information DB 56 and the patient's pain prediction information obtained from the learning model 61. In this case, the pain treatment auxiliary information may be generated based on the patient's pain-related information, diagnosis and treatment information, and literature information in addition to the past pain pattern and the pain prediction information.

[0100] The control unit 51 stores the generated pain treatment auxiliary information in the pain treatment auxiliary information DB 56 (database) ( S18 ), outputs the generated pain treatment auxiliary information ( S19 ), and ends the processing.

[0101] Fig.17 is a diagram showing an example of the generation order of the learning model 61. The control unit 51 obtains training data including pain-related information, diagnosis and treatment information, pain prediction information, and literature information related to pain treatment for multiple patients (S31). Based on the obtained training data, the control unit 51 generates the learning model 61 so that when the pain-related information, diagnosis and treatment information, and literature information related to pain treatment for each patient are input, the prediction information of the pain of each patient is output (S32). The control unit 51 stores the generated learning model 61 in the storage unit 59 (S33), and ends the processing.

[0102] Description of Reference Numerals

[0103] 1…communication network; 10…patient terminal device; 20…equipment; 30…assistant terminal device; 40…doctor terminal device; 50…pain treatment assistance device; 51…control unit; 52…communication unit; 53…storage; 54…pain-related information DB; 55…medical information DB; 56…pain treatment assistance information DB; 57…literature information DB; 58…auxiliary information generation unit; 59…storage unit; 60…computer program; 61…learning model.

Claims

1. A computer program, characterized in that Causes the computer to perform the following processing: Obtain pain-related information related to the patient's pain, inputting the acquired pain-related information into a learning model that outputs predicted information of pain when pain-related information is input, thereby acquiring predicted information of the patient's pain, Based on the acquired prediction information of pain, pain treatment support information for the patient is generated, and the generated pain treatment support information is output.

2. The computer program according to claim 1, characterized in that The pain-related information includes at least one of an evaluation index for evaluating pain, vital data, medication information, and physical condition.

3. The computer program according to claim 1, characterized in that Causes the computer to perform the following processing: obtaining medical information related to the medical treatment of the patient, Furthermore, the acquired medical information is input to the learning model that outputs prediction information of pain when medical information is input, thereby acquiring prediction information of pain of the patient.

4. The computer program according to claim 3, characterized in that The diagnosis and treatment information includes at least one of diagnosis information, examination information, treatment information, prescription information and disease prognosis prediction information.

5. The computer program according to claim 1, characterized in that Causes the computer to perform the following processing: Obtain literature information related to pain treatment, Furthermore, the acquired literature information is input to the learning model that outputs prediction information of pain when literature information related to pain treatment is input, thereby acquiring prediction information of pain of the patient.

6. The computer program according to any one of claims 1 to 5, characterized in that The pain treatment assistance information includes at least one of recommended prescription information and recommended treatment information for assisting a doctor examining the patient.

7. The computer program according to any one of claims 1 to 5, characterized in that The pain treatment auxiliary information includes: coping methods for assisting the patient.

8. The computer program according to any one of claims 1 to 5, characterized in that Causes the computer to perform the following processing: A predicted pain pattern indicating a temporal transition of a pain degree based on the predicted information on the pain of the patient is output.

9. An information processing device, characterized in that: have: a first acquisition unit that acquires pain-related information related to the patient's pain; a second acquisition unit that inputs the acquired pain-related information into a learning model that outputs prediction information of pain when pain-related information is input, thereby acquiring prediction information of pain of the patient; a generating unit that generates pain treatment auxiliary information for the patient based on the acquired pain prediction information; as well as An output unit outputs the generated pain treatment auxiliary information.

10. An information processing method, characterized in that: Obtain pain-related information related to the patient's pain, inputting the acquired pain-related information into a learning model that outputs predicted information of pain when pain-related information is input, thereby acquiring predicted information of the patient's pain, Based on the acquired prediction information of pain, pain treatment support information for the patient is generated, and the generated pain treatment support information is output.

11. A learning model generation method, characterized in that: obtaining first training data, the first training data comprising: pain-related information related to pain of a plurality of patients, and prediction information of pain of the plurality of patients, Based on the acquired first training data, a learning model is generated so as to output prediction information of pain when pain-related information is input.

12. The learning model generation method according to claim 11, characterized in that: obtaining second training data, the second training data further comprising: diagnosis and treatment information of a plurality of patients, literature information related to pain treatment, and prediction information of pain of the plurality of patients, The learning model is generated based on the acquired second training data so as to output prediction information of pain when medical treatment information and literature information related to pain treatment are input.

Images