System for assisting a medical and pharmaceutical information personnel in charge

Abstract

One non-limiting and exemplary embodiment provides a system for supporting medical and pharmaceutical information personnel in charge (MR).SOLUTION: A) information of patients having a likelihood of suffering from Fabry's disease, and / or b) information of physicians having a likelihood of being able to diagnose and / or treat Fabry's disease to a medical and pharmaceutical information personnel in charge. Also provided are a method, apparatus, and program for predicting the likelihood that a patient is suffering from Fabry disease, and a method, apparatus, and program for predicting the likelihood that a doctor can diagnose and / or treat Fabry disease.SELECTED DRAWING: None

Description

[Technical Field]

[0001] The present disclosure relates to a system for assisting medical representatives (MRs), and more specifically to a system that provides a medical representative with a) information on patients who may have Fabry disease and / or b) information on doctors who may be able to diagnose and / or treat Fabry disease. The present disclosure also relates to a method for predicting the likelihood that a patient has Fabry disease, a computer program for predicting the likelihood that a patient has Fabry disease, a method for generating a trained machine learning model for use in predicting the likelihood that a patient has Fabry disease, an apparatus for predicting the likelihood that a patient has Fabry disease, a method for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease, a computer program for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease, a method for generating a trained machine learning model for use in predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease, and an apparatus for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease. [Background technology]

[0002] Fabry disease, also known as Anderson-Fabry disease or Fabry disease, is a lysosomal storage disorder (designated as an intractable disease in Japan). It is an X-linked recessive genetic disorder caused by a deficiency or reduced activity of α-galactosidase A (α-galA, α-GalA), an intracellular lysosomal enzyme, resulting in abnormalities in intracellular glycolipid metabolism. Due to its X-linked inheritance pattern, it is more prevalent in men. While women may also develop the disease, their symptoms generally tend to be milder than men's, and range from asymptomatic to severe. The disease may be detected in early childhood due to symptoms such as sharp pain in the hands and feet, anhidrosis, and a rash on the buttocks and genitals. However, delayed detection or lack of appropriate treatment can lead to the onset and progression of renal, cardiac, and cerebral symptoms during adolescence and middle age. Alpha-galactosidase enzyme preparations are available as treatments for Fabry disease. Patent documents 1 to 3 disclose treatments for Fabry disease. Fabry disease is known to be a rare disease.

[0003] Because rare diseases such as Fabry disease only affect a small number of patients, doctors tend to have little experience diagnosing them, which inevitably leads to a certain number of doctors being unable to properly diagnose patients with Fabry disease. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2023-052005 [Patent Document 2] Japanese Patent Publication No. 2023-027053 [Patent Document 3] Japanese Patent Publication No. 2022-130589 Summary of the Invention [Problem to be solved by the invention]

[0005] If symptoms of Fabry disease progress, cardiac pacemakers, bypass surgery, hemodialysis, or kidney transplants may be necessary, so it is important to start treatment as early as possible. However, the signs, symptoms, and progression of Fabry disease are diverse, making it difficult to diagnose in childhood; it is often diagnosed in adulthood when cardiac or renal abnormalities are present. Newborn screening studies suggest that the incidence of Fabry disease is approximately 1 in 7,000, while the number of Fabry disease patients receiving treatment in Japan is estimated to be approximately 1,000, meaning that there are an estimated 16,000 potential patients.

[0006] As such, it is believed that there are a certain number of patients who have not been properly diagnosed with Fabry disease, and therefore it is necessary to support doctors and hospitals so that they can properly diagnose patients with Fabry disease and start appropriate treatment. In supporting doctors and medical facilities, it is important for medical representatives (MRs) to provide doctors with appropriate information, and a system to support such activities could be useful. [Means for solving the problem]

[0007] The present disclosure provides a system for supporting medical representatives (MRs), more specifically, a system that provides a medical representative with a) information on patients who may have Fabry disease and / or b) information on doctors who may be able to diagnose and / or treat Fabry disease. The present disclosure also provides a method for predicting the likelihood that a patient has Fabry disease, a computer program for predicting the likelihood that a patient has Fabry disease, a method for generating a trained machine learning model for use in predicting the likelihood that a patient has Fabry disease, an apparatus for predicting the likelihood that a patient has Fabry disease, a method for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease, a computer program for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease, a method for generating a trained machine learning model for use in predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease, and an apparatus for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease.

[0008] The present disclosure provides the following as more specific examples. [Embodiment 1] A system for supporting medical representatives (MRs), A server having a communication unit for connecting to a user terminal and a storage unit storing information on patients and / or doctors; A user terminal having a communication unit for connecting to a server and an output unit for providing information to a user. and stored on the server a) information on individuals who may have Fabry disease; and / or b) A system that provides medical representatives with information on doctors who may be able to diagnose and / or treat Fabry disease via output to a user terminal. [Embodiment 2] Stored on the server a) information on individuals who may have Fabry disease; and b) A system as described in embodiment 1, which provides information on doctors who may be able to diagnose and / or treat Fabry disease to medical representatives via output to a user terminal. [Embodiment 3] A system as described in embodiment 1, wherein the patient information includes information on the area where the hospital where the patient may be treated is located. [Embodiment 4] 2. The system of embodiment 1, wherein the patient information includes a number of patients likely to have Fabry disease associated with a particular region. [Embodiment 5] 2. The system of embodiment 1, wherein information about areas with a high number of patients who may have Fabry disease is provided to medical representatives in preference to areas with a low number of patients. [Embodiment 6] A system as described in embodiment 1, wherein the patient information includes a list of one or more hospitals where the patient may be treated. [Embodiment 7] A system as described in embodiment 1, which provides medical representatives with information about doctors affiliated with hospitals where patients who may have Fabry disease may visit and who may be able to diagnose and / or treat Fabry disease. [Embodiment 8] 2. The system of embodiment 1, wherein the patient information includes a value indicating the likelihood that the patient has Fabry disease. [Embodiment 9] The system of embodiment 8, wherein the value indicating the possibility that a patient has Fabry disease is a predicted value by a prediction model or a stratified predicted value. [Embodiment 10] The system of embodiment 9, wherein the predictive model is a machine learning model trained to input at least a portion of a patient's real-world data, including medical receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from wearable devices, and output a value indicating the likelihood that the patient has Fabry disease. [Embodiment 11] The system of embodiment 9, wherein the predictive model is a machine learning model trained to output a value indicating the likelihood that a patient has Fabry disease in response to input of one or more explanatory variables including a patient's diagnostic history, testing history, prescription history, medical treatment history, and / or attributes. [Embodiment 12] The system of embodiment 11, wherein the predictive model is a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients diagnosed by a physician as having or suspected of having Fabry disease. [Embodiment 13] The system of embodiment 11, wherein the predictive model is a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients who have been diagnosed by a physician as not having or not suspected of having Fabry disease. [Embodiment 14] The system of embodiment 11, wherein the diagnosis in the history of diagnoses includes a diagnosis that confirms the disease. [Embodiment 15] The system of embodiment 11, wherein a diagnosis in the diagnostic history includes a diagnosis indicating a suspicion of disease. [Embodiment 16] The diagnosis in the diagnostic history was disseminated intravascular coagulation, type 2 diabetes, congestive heart failure, Sjögren's syndrome, rheumatoid arthritis, acute myocardial infarction, acute progressive glomerulonephritis, rapidly progressive glomerulonephritis, angina pectoris, blood coagulation disorder, hypothyroidism, hyperthyroidism, valvular heart disease, heart failure, deep vein thrombosis, decreased renal function, systemic lupus erythematosus, iron deficiency anemia, diabetes, urinary tract infection, cerebral infarction, pneumonia, irregular heartbeat, or urinary tract infection. The system of embodiment 11 includes diagnosing a disease selected from the group consisting of: pulmonary artery disease, collagen disease, mitral valve regurgitation, type 2 diabetes with or without diabetic complications, congestive heart failure, suspected disease number, angina pectoris, hypothyroidism, hyperphosphatemia, heart failure, renal anemia, sleep apnea syndrome, amplified valvular regurgitation, dehydration, hypertrophic cardiomyopathy, arrhythmia, peripheral neuropathy, cholinergic urticaria, plasma cell myeloma, chronic heart failure, and chronic renal failure. [Embodiment 17] The system of embodiment 11, wherein the tests in the test history include a creatinine test. [Embodiment 18] The system of embodiment 11, wherein the attributes include attributes selected from the group consisting of gender, age, age at diagnosis, and number of suspected diseases. [Embodiment 19] 12. The system of embodiment 11, wherein the machine learning model is selected from the group consisting of decision trees, random forests, LightGBM, stacking, logistic regression, lasso regression, support vector machines, multilayer perceptrons, neural networks, and combinations thereof. [Embodiment 20] A system as described in embodiment 1, wherein the doctor's information includes a list of one or more hospitals to which the doctor is affiliated. [Embodiment 21] 2. The system of embodiment 1, wherein the physician information comprises a value indicating the likelihood that the physician will be able to diagnose and / or treat Fabry disease. [Embodiment 22] The system described in embodiment 21, wherein the value indicating the likelihood that a physician can diagnose and / or treat Fabry disease is determined based on one or more items including the physician's gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers on Fabry disease. [Embodiment 23] 23. The system of embodiment 22, wherein a value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease is determined based on the number or frequency of accesses to digital content related to Fabry disease. [Embodiment 24] The system of embodiment 21, wherein the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease is a predicted value by a prediction model or a stratified value of the predicted value. [Embodiment 25] The system described in embodiment 24, wherein the predictive model is a machine learning model trained to output a value indicating the likelihood that a doctor will be able to diagnose and / or treat Fabry disease in response to input of one or more explanatory variables including the doctor's gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers on Fabry disease. [Embodiment 26] The system of embodiment 25, wherein the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, access behavior history of digital content related to Fabry disease, purchase history of medications related to Fabry disease, and / or history of writing papers on Fabry disease of physicians known to be able to diagnose and / or treat Fabry disease. [Embodiment 27] The system of embodiment 25, wherein the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, access behavior history of digital content related to Fabry disease, purchase history of medications related to Fabry disease, and / or history of writing papers on Fabry disease of doctors known to be unable to diagnose and / or treat Fabry disease. [Embodiment 28] 26. The system of embodiment 25, wherein the machine learning model is selected from the group consisting of decision trees, random forests, LightGBM, stacking, logistic regression, lasso regression, support vector machines, multilayer perceptrons, neural networks, and combinations thereof. [Embodiment 29] 29. The system of any one of embodiments 1 to 28, which provides a medical representative with information about patients and / or doctors in the medical representative's area of ​​responsibility. [Embodiment 30] A system as described in any one of embodiments 1 to 28, which suggests to a medical representative which hospitals and / or doctors the medical representative should contact. [Embodiment 31] The system of embodiment 30, wherein the hospital that the medical representative should contact is determined based on the number of patients who may have Fabry disease and who may be visiting the hospital. [Embodiment 32] The system of embodiment 30, wherein the hospital that the medical representative should contact is determined based on the number of doctors affiliated with that hospital who have the potential to diagnose and / or treat Fabry disease. [Embodiment 33] 1. A method for predicting the likelihood that a patient has Fabry disease, comprising: Inputting at least a portion of the patient's real-world data, including receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from a wearable device, into a predictive model; and obtaining an output from the predictive model indicating the likelihood that the patient has Fabry disease. A method comprising: [Embodiment 34] The method of embodiment 33, wherein the value indicating the possibility that the patient has Fabry disease is a predicted value by a prediction model or a stratified predicted value. [Embodiment 35] The method of embodiment 33, wherein the predictive model is a machine learning model trained to input at least a portion of the patient's real-world data, including medical receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from wearable devices, and output a value indicating the likelihood that the patient has Fabry disease. [Embodiment 36] The method of embodiment 33, wherein the predictive model is a machine learning model trained to output a value indicating the likelihood that a patient has Fabry disease in response to input of one or more explanatory variables including a patient's diagnostic history, test history, prescription history, medical treatment history, and / or attributes. [Embodiment 37] The method of embodiment 36, wherein the predictive model is a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients diagnosed by a physician as having or suspected of having Fabry disease. [Embodiment 38] The method of embodiment 36, wherein the predictive model is a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients who have been diagnosed by a physician as not having or not suspected of having Fabry disease. [Embodiment 39] 37. The method of embodiment 36, wherein the diagnoses in the history of diagnoses include diagnoses that confirm the disease. [Embodiment 40] 37. The method of embodiment 36, wherein the diagnosis in the diagnostic history includes a diagnosis indicating suspicion of disease. [Embodiment 41] The diagnosis in the diagnostic history was disseminated intravascular coagulation, type 2 diabetes, congestive heart failure, Sjögren's syndrome, rheumatoid arthritis, acute myocardial infarction, acute progressive glomerulonephritis, rapidly progressive glomerulonephritis, angina pectoris, blood coagulation disorder, hypothyroidism, hyperthyroidism, valvular heart disease, heart failure, deep vein thrombosis, decreased renal function, systemic lupus erythematosus, iron deficiency anemia, diabetes, urinary tract infection, cerebral infarction, pneumonia, or urinary tract infection. 37. The method of embodiment 36, comprising diagnosing a disease selected from the group consisting of arrhythmia, collagen disease, mitral valve regurgitation, type 2 diabetes with or without diabetic complications, congestive heart failure, suspected disease number, angina pectoris, hypothyroidism, hyperphosphatemia, heart failure, renal anemia, sleep apnea syndrome, amplified valvular regurgitation, dehydration, hypertrophic cardiomyopathy, arrhythmia, peripheral neuropathy, cholinergic urticaria, plasma cell myeloma, chronic heart failure, and chronic renal failure. [Embodiment 42] 37. The method of embodiment 36, wherein the tests in the testing history include a creatinine test. [Embodiment 43] 37. The method of embodiment 36, wherein the attributes include attributes selected from the group consisting of sex, age, age at diagnosis, and number of suspected diseases. [Embodiment 44] 37. The method of embodiment 36, wherein the machine learning model is selected from the group consisting of a decision tree, a random forest, a LightGBM, stacking, logistic regression, lasso regression, a support vector machine, a multilayer perceptron, a neural network, and combinations thereof. [Embodiment 45] A method for identifying hospitals where patients with Fabry disease may be treated, comprising the steps of: identifying patients who are predicted to have a possibility of having Fabry disease by a method described in any one of embodiments 33 to 44; identifying the area where the hospital where the patient is treated is located from the attribute information of the identified patients; and identifying hospitals located in the identified area. [Embodiment 46] 1. A computer program for predicting the likelihood that a patient has Fabry disease, the computer program comprising, when implemented on a computer: Inputting at least a portion of the patient's real-world data, including receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from a wearable device, into a predictive model; and obtaining an output from the predictive model indicating the likelihood that the patient has Fabry disease. A program containing instructions that cause a processor to execute the program. [Embodiment 47] A computer program for causing a computer to execute the method according to any one of embodiments 33 to 44. [Embodiment 48] A computer-readable recording medium having recorded thereon the computer program described in embodiment 47. [Embodiment 49] A trained machine learning model used to cause a computer to execute the method described in any one of embodiments 35 to 44. [Embodiment 50] A method for generating a trained machine learning model for use in predicting the likelihood that a patient has Fabry disease, the method using teacher data for training, with at least a portion of the patient's real-world data, including medical receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from wearable devices, as explanatory variables, and the presence or absence of Fabry disease or suspicion of Fabry disease as the objective variable. [Embodiment 51] A method for generating a trained machine learning model as described in embodiment 50, wherein at least a portion of the patient's real-world data includes the patient's diagnosis history, examination history, prescription history, medical procedure history, and / or attributes. [Embodiment 52] 1. A device for predicting the likelihood that a patient has Fabry disease, comprising: an input unit for inputting at least a portion of the patient's real-world data, including receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data derived from a wearable device, into the prediction model; a prediction portion for predicting the likelihood that a patient has Fabry disease using a predictive model; and an output section for outputting a value indicating the likelihood that the predicted patient has Fabry disease; An apparatus having: [Embodiment 53] An apparatus comprising means for carrying out the method according to any one of embodiments 33 to 44. [Embodiment 54] An apparatus comprising a processor connected to a storage device storing a computer program according to embodiment 47, the processor being capable of executing instructions of the program. [Embodiment 55] 50. An apparatus for predicting the likelihood that a patient has Fabry disease, comprising a prediction unit that makes predictions using the trained machine learning model described in embodiment 49. [Embodiment 56] 1. A method for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease, comprising: A step of inputting data including the doctor's gender, age, place of origin, alma mater, year of graduation, work facility history, medical department history, access behavior history of digital content related to Fabry disease, purchase history of medication related to Fabry disease, and / or paper writing history related to Fabry disease into a prediction model; and obtaining an output from the predictive model that indicates the likelihood that a physician will be able to diagnose and / or treat Fabry disease. A method comprising: [Embodiment 57] 57. The method of embodiment 56, wherein the predictive model makes a prediction based on the number or frequency of accesses to digital content related to Fabry disease. [Embodiment 58] The method of embodiment 56, wherein the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease is a predicted value by a prediction model or a stratified predicted value. [Embodiment 59] The method described in embodiment 58, wherein the predictive model is a machine learning model trained to output a value indicating the likelihood that the doctor will be able to diagnose and / or treat Fabry disease in response to input of one or more explanatory variables including the doctor's gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers on Fabry disease. [Embodiment 60] The method of embodiment 59, wherein the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department affiliation, access behavior history of digital content related to Fabry disease, purchase history of medications related to Fabry disease, and / or history of writing papers on Fabry disease of physicians known to be able to diagnose and / or treat Fabry disease. [Embodiment 61] The method of embodiment 59, wherein the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers on Fabry disease of physicians known to be unable to diagnose and / or treat Fabry disease. [Embodiment 62] 60. The method of embodiment 59, wherein the machine learning model is selected from the group consisting of decision trees, random forests, LightGBM, stacking, logistic regression, lasso regression, support vector machines, multilayer perceptrons, neural networks, and combinations thereof. [Embodiment 63] A method for identifying a hospital where a doctor who can diagnose and / or treat Fabry disease may be affiliated, the method comprising the steps of identifying a doctor who is predicted to have the potential to diagnose and / or treat Fabry disease by a method described in any one of embodiments 56 to 62, and identifying the hospital to which the doctor is affiliated from the attribute information of the identified doctor. [Embodiment 64] 1. A computer program that predicts the likelihood that a physician will be able to diagnose and / or treat Fabry disease, which, when implemented on a computer, comprises: A step of inputting data including the doctor's gender, age, place of origin, alma mater, year of graduation, work facility history, medical department history, access behavior history of digital content related to Fabry disease, purchase history of medication related to Fabry disease, and / or paper writing history related to Fabry disease into a prediction model; and obtaining an output from the predictive model that indicates the likelihood that a physician will be able to diagnose and / or treat Fabry disease. A program containing instructions that cause a processor to execute the program. [Embodiment 65] 65. The program of embodiment 64, wherein the predictive model makes a prediction based on the number or frequency of accesses to digital content related to Fabry disease. [Embodiment 66] A computer program for causing a computer to execute the method according to any one of embodiments 56 to 62. [Embodiment 67] A computer-readable recording medium having recorded thereon the computer program described in embodiment 66. [Embodiment 68] A trained machine learning model used to cause a computer to execute the method described in any one of embodiments 59 to 62. [Embodiment 69] A method for generating a trained machine learning model to be used to predict the likelihood that a doctor will be able to diagnose and / or treat Fabry disease, which uses training data to train a trained machine learning model, with the doctor's gender, age, place of origin, alma mater, year of graduation, history of working facility, history of medical department, history of access to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers on Fabry disease as explanatory variables, and whether or not the doctor will be able to diagnose and / or treat Fabry disease as the objective variable. [Embodiment 70] 1. A device for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease, comprising: an input section for inputting data into the prediction model, including the doctor's gender, age, place of origin, alma mater, year of graduation, history of facility where the doctor works, history of medical department to which the doctor belongs, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers related to Fabry disease; a prediction component for predicting the likelihood of a physician being able to diagnose and / or treat Fabry disease using a predictive model; and an output unit for outputting a value indicative of the predicted likelihood that the physician will be able to diagnose and / or treat Fabry disease; An apparatus having: [Embodiment 71] 71. The device of embodiment 70, wherein the predictive model makes predictions based on the number or frequency of accesses to digital content related to Fabry disease. [Embodiment 72] An apparatus comprising means for carrying out the method described in any one of embodiments 56 to 62. [Embodiment 73] An apparatus comprising a processor connected to a memory device storing a computer program as described in embodiment 64, and capable of executing instructions of the program. [Embodiment 74] An apparatus for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease, comprising a prediction unit that makes predictions using the trained machine learning model described in embodiment 68. [Effects of the Invention]

[0009] The system for supporting medical representatives (MRs) according to the present disclosure can, for example, identify patients who may have Fabry disease and doctors who may be able to diagnose and / or treat Fabry disease, and present this information to medical representatives, thereby helping the medical representatives to more efficiently contact appropriate doctors and / or hospitals and provide appropriate medical information. As a result, it is expected that the likelihood that undiagnosed patients with Fabry disease will be increased and treatment for the disease will be able to be started earlier.

[0010] Furthermore, the method or program disclosed herein for predicting the likelihood that a patient has Fabry disease can also be used to predict, for example, whether or not there are patients with Fabry disease in a particular area. By combining this with information about the area where doctors or hospitals are located, it is possible to identify hospitals where patients with Fabry disease are likely to visit and doctors who are likely to examine patients with Fabry disease. This allows medical representatives and others to narrow down the doctors or hospitals that are suitable to provide information about therapeutic drugs for Fabry disease.

[0011] Furthermore, the method or program disclosed herein for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease can also be useful, for example, for determining whether a particular doctor is suitable for diagnosing a rare disease. By combining this with information about the area in which the doctor or hospital is located and the results of predicting whether there are patients with Fabry disease in a particular area, it is possible for MRs and others to narrow down the doctors or hospitals that are suitable to provide information about therapeutic drugs for Fabry disease and the like. [Brief explanation of the drawings]

[0012] [Figure 1]1 shows a schematic configuration of an exemplary computer that can be used to implement aspects of the present disclosure. The exemplary computer (100) includes a control unit (101), a memory unit (102), a peripheral device I / F unit (103), an input unit (104), a display unit (105), a communication unit (106), and a bus (110). The computer (100) can be connected to an external server (130) and a database (140) via a network (120). [Figure 2] FIG. 2 is a schematic diagram of an exemplary system that can be used to implement aspects of the present disclosure. The exemplary system includes a server (210) and a user terminal (220), with the server and user terminal connected via a network (230). Although not shown here, multiple user terminals can be connected to the server. The exemplary server includes components such as a memory unit (214) and a communication unit (216), and has a control unit (212) that controls these components. The exemplary user terminal includes an output unit (224) and a communication unit (226), and has a control unit (222) that controls these components. The server may further be connected to a patient prediction device (240) that predicts the likelihood that a patient has Fabry disease and a doctor prediction device (250) that predicts the likelihood that a doctor will be able to diagnose and / or treat Fabry disease. Prediction results calculated by these prediction devices may be stored in the server's memory. [Figure 3] 3 is a schematic diagram of an exemplary device for predicting the likelihood that a patient has Fabry disease (also referred to herein as a patient prediction device) that can be used to implement aspects of the present disclosure. The exemplary patient prediction device (300) has an input unit (310), a prediction unit (320), and an output unit (230). [Figure 4] 4 is a schematic diagram of an exemplary device for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease (also referred to herein as a physician prediction device) that may be used to implement aspects of the present disclosure. The exemplary physician prediction device (400) includes an input unit (410), a prediction unit (420), and an output unit (430). [Figure 5]FIG. 5 shows an exemplary flow chart of a computer program for predicting the likelihood that a patient has Fabry disease that can be used to implement aspects of the present disclosure. [Figure 6] FIG. 6 shows an exemplary flowchart of a computer program for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease that may be used to implement aspects of the present disclosure. [Figure 7] Figure 7 shows the AUC of the random forest (RF) model among the generated trained machine learning models. [Figure 8] Figure 8 shows the results of analyzing the feature importance in a trained random forest (RF) model. [Figure 9] Figure 9 shows the AUC of the LightGBM (LGBM) model among the generated trained machine learning models. [Figure 10] Figure 10 shows the results of analyzing the importance of features in the trained LightGBM (LGBM) model. [Figure 11] Figure 11 shows the AUC of the logistic regression (LR) model among the generated trained machine learning models. [Figure 12] Figure 12 shows the results of analyzing the importance of features in a trained logistic regression (LR) model. DETAILED DESCRIPTION OF THE INVENTION

[0013] A system to support medical representatives (MRs) In one aspect, the present disclosure relates to a system for supporting medical representatives (MRs), who are individuals whose main job is to provide, collect, and communicate information on the quality, efficacy, safety, etc. of pharmaceuticals by visiting healthcare professionals to ensure the proper use of pharmaceuticals.

[0014] In some embodiments, the system of the present disclosure includes a server having a communication unit for connecting to a user terminal and a memory unit storing patient and / or doctor information, and a user terminal having a communication unit for connecting to the server and an output unit for providing information to a user. In some embodiments, the memory unit of the server included in the system of the present disclosure stores a) information on patients who may have Fabry disease and / or b) information on doctors who may be able to diagnose and / or treat Fabry disease. In some embodiments, the system of the present disclosure can provide the information stored in the server to a medical representative via output to the user terminal. In some embodiments, the system of the present disclosure may include one or more servers and one or more user terminals. In some embodiments, multiple user terminals may be connected to the server of the present disclosure. In some embodiments, the server of the present disclosure may be connected to a patient prediction device that predicts the likelihood that a patient has Fabry disease and / or a doctor prediction device that predicts the likelihood that a doctor will be able to diagnose and / or treat Fabry disease. Thus, in some embodiments, the systems of the present disclosure may include a patient prediction device that predicts the likelihood that a patient has Fabry disease and / or a physician prediction device that predicts the likelihood that a physician will be able to diagnose and / or treat Fabry disease.

[0015] In some embodiments, the system of the present disclosure may provide a medical representative with a combination of a) information on patients who may have Fabry disease and b) information on doctors who may be able to diagnose and / or treat Fabry disease. Thus, in some embodiments, the system may provide a medical representative with information on doctors who may be able to diagnose and / or treat Fabry disease and who belong to hospitals where patients who may have Fabry disease may be visiting. In this way, by combining and using a) information on patients who may have Fabry disease and b) information on doctors who may be able to diagnose and / or treat Fabry disease, it is possible to more efficiently support the activities of medical representatives.

[0016] In some embodiments, the information stored on the server may be updated daily, weekly, or monthly.

[0017] In some embodiments, the server includes a controller including a processor, such as a central processing unit (CPU), and memory, such as ROM and RAM. In some embodiments, the server may be a cloud server distributed over a network.

[0018] In some embodiments, the communication unit of the server is connected to the user terminal via a wired or wireless network such as the Internet, a LAN, or a WAN. In some embodiments, the storage unit of the server may be a storage device such as, but not limited to, a hard disk, an SSD, or a flash memory. In some embodiments, the storage unit of the server may be cloud storage.

[0019] In some embodiments, the user terminal may be a device such as, but not limited to, a smartphone, a tablet, a smartwatch, a laptop, a desktop PC, etc. In some embodiments, the user terminal includes a controller including a processor, such as a central processing unit (CPU), and memory, such as ROM and RAM.

[0020] In some embodiments, the communication unit of the user terminal is connected to the server via a wired or wireless network such as the Internet, a LAN, or a WAN. In some embodiments, the user terminal may also include a memory unit that stores information, such as information obtained from the server of the present disclosure. In some embodiments, the memory unit of the user terminal may be a storage device such as a hard disk, an SSD, or a flash memory, but is not limited to these. In some embodiments, the memory unit of the user terminal may be cloud storage. In some embodiments, the output unit of the user terminal may be an output device such as a display, a printer, or a speaker, but is not limited to these.

[0021] Information on patients who may have Fabry disease In some embodiments, the server of the present disclosure stores information about patients who may have Fabry disease in a storage unit. In some embodiments, the patient information includes a list of one or more hospitals where the patient may be treated. The hospitals where the patient may be treated may be identified from among hospitals located in a region based on regional information obtained, for example, from medical receipt data of patients who may have Fabry disease. If there are multiple hospitals in a region, a list of hospitals may be created. The hospitals included in the list may be ranked based on the type and number of medical departments, the presence or absence of doctors with expertise in specific fields, etc. The patient information may also include regional information about the locations of hospitals where the patient may be treated. Thus, in some embodiments, the system of the present disclosure may provide a medical representative with the names of regions where hospitals where the patient may be treated are located. The patient information may also include the number of patients who may have Fabry disease associated with a particular region. Thus, the system of the present disclosure may present to the medical representative regions and / or hospitals where there are thought to be more potential Fabry disease patients than others. The system of the present disclosure may provide medical representatives with information about regions with a high number of patients who may have Fabry disease in priority over regions with a low number of patients. In some embodiments, for example, the system of the present disclosure may present medical representatives with regions with a high number of patients in order.

[0022] In some embodiments, the patient information includes a value indicating the likelihood that the patient has Fabry disease. In some embodiments, the value indicating the likelihood that the patient has Fabry disease may be, for example, but not limited to, an integer or real number between 0 and 100. In some embodiments, the value indicating the likelihood that the patient has Fabry disease may have any threshold value; for example, only patients with a value of 50 or greater, 60 or greater, 70 or greater, or 80 or greater may be considered potential Fabry disease patients. In some embodiments, the value indicating the likelihood that the patient has Fabry disease may be stratified into, for example, two, three, four, five, six, seven, eight, nine, or ten levels. For example, in the case of a two-level setting, the value indicating the likelihood that the patient has Fabry disease may be stratified into two levels: a "low likelihood" and a "high likelihood." For example, in the case of a 3-point scale, the value indicating the likelihood that a patient has Fabry disease may be stratified into three categories: "high," "moderate," or "low." For example, in the case of a 10-point scale, the value indicating the likelihood that a patient has Fabry disease may be ranked using numbers from 1 to 10.

[0023] In some embodiments, the value indicating the likelihood that a patient has Fabry disease is a predicted value or a stratified predicted value obtained by a predictive model, which may be, but is not limited to, a machine learning model, a statistical model, a simulation model, a rule-based model, a time series model, or a symbolic regression model.

[0024] In some embodiments, the predictive model may output a value indicating the possibility that a patient has Fabry disease, for example, but not limited to, a real number between 0 and 1, or an integer or real number between 0 and 100. In some embodiments, the value indicating the possibility that a patient has Fabry disease may be set to any threshold value; for example, only patients having a value of 0.5 or more, 0.6 or more, 0.7 or more, or 0.8 or more, or a value of 50 or more, 60 or more, 70 or more, or 80 or more may be considered to be potential Fabry disease patients, and other values ​​may be set to 0. In some embodiments, the value indicating the possibility that a patient has Fabry disease may be stratified into, for example, 2 levels, 3 levels, 4 levels, 5 levels, 6 levels, 7 levels, 8 levels, 9 levels, or 10 levels. For example, in the case of a 2-level setting, the value indicating the possibility that a patient has Fabry disease may be stratified into two levels: "low possibility" and "high possibility." For example, in the case of a 3-point scale, the value indicating the likelihood that a patient has Fabry disease may be stratified into three categories: "high," "moderate," or "low." For example, in the case of a 10-point scale, the value indicating the likelihood that a patient has Fabry disease may be ranked using numbers from 1 to 10.

[0025] In some embodiments, the predictive model is a machine learning model trained to take as input at least a portion of a patient's real-world data, including medical receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from wearable devices, and output a value indicating the likelihood that the patient has Fabry disease.

[0026] A machine learning model is a mathematical model that learns patterns and relationships from data and then uses that learning to make predictions or decisions about new data. There are several main types of machine learning models: 1. Supervised learning: The model is given input data and corresponding labels and learns the relationship between them. There are two types of supervised learning: classification and regression. In classification, the model classifies input data into specific classes, while in regression, the model predicts continuous values. 2. Unsupervised learning: Techniques for finding patterns and structure in unlabeled datasets, including clustering and dimensionality reduction. Clustering involves grouping similar data together, while dimensionality reduction involves reducing the dimensionality of the data and extracting important features. 3. Reinforcement learning: An agent learns through interaction with the environment, learning actions that maximize rewards. This is widely used in areas such as game playing and robot control. 4. Semi-supervised learning: A combination of supervised and unsupervised learning, used when only part of the data is labeled. 5. Transfer learning: A technique for transferring knowledge learned in one task to other related tasks, using a pre-trained model and adapting it to the new task.

[0027] In some embodiments according to the present disclosure, the machine learning model is generated by supervised learning. In some embodiments, the machine learning model is a classification model, e.g., classifying a patient as having or not having Fabry disease. In some embodiments, the machine learning model is a regression model, e.g., outputting a continuous value indicating the likelihood that a patient has Fabry disease.

[0028] In some embodiments, the input data (also called explanatory variables or features) can be real-world patient data, including medical receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from wearable devices.

[0029] In the healthcare context, "real-world data (RWD)" generally refers to data collected from actual clinical practice or healthcare activities outside of clinical trials or research. Specifically, it includes patient medical history, diagnostic information, treatment history, drug use, clinical results, health outcomes, and living environment. Real-world data can be an important source of information when evaluating a patient's clinical progress and treatment effectiveness.

[0030] Examples of the number of patients included in the input data include values ​​of 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 200 or more, 300 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more, 2000 or more, 3000 or more, 4000 or more, 5000 or more, 6000 or more, 7000 or more, 8000 or more, 9000 or more, 10,000 or more, or 100,000 or more.

[0031] The number of types of explanatory variables (or features) included in the input data can be, for example, 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, 100 or more, 200 or more, 300 or more, 400 or more, 500 or more, 600 or more, 700 or more, 800 or more, 900 or more, 1000 or more, 2000 or more, 3000 or more, 4000 or more, 5000 or more, 6000 or more, 7000 or more, 8000 or more, 9000 or more, 10000 or more, or 10 Examples of the number include 0000 or more, and 1,000,000 or less, 100,000 or less, 10,000 or less, 9,000 or less, 8,000 or less, 7,000 or less, 6,000 or less, 5,000 or less, 4,000 or less, 3,000 or less, 2,000 or less, 1,000 or less, 900 or less, 800 or less, 700 or less, 600 or less, 500 or less, 400 or less, 300 or less, 200 or less, 100 or less, 90 or less, 80 or less, 70 or less, 60 or less, 50 or less, 40 or less, 30 or less, or 20 or less.

[0032] Medical receipt (medical fee statement) data refers to data that records information about medical treatment and drug prescriptions received from patients by medical institutions. This data is mainly used for insurance claims. Medical receipt data includes the patient's name, date of birth, gender, medical specialty, prescribed drugs and treatment, consultation date and time, and information about the medical institution. This information is used for medical fee claims, statistical analysis, and medical policy planning. Anonymized medical receipt information can be obtained, for example, from the Anonymous Medical Insurance Related Information Database (NDB) run by the Ministry of Health, Labor and Welfare of Japan. Examples of medical receipt databases available as products include the JMDC Claims Database, an epidemiological medical receipt database from JMDC Co., Ltd., and the Cross Fact Pharmaceutical Receipt Database and Cross Fact Social Insurance Receipt Database, integrated medical databases from INTAGE Real World Co., Ltd.

[0033] Diagnosis Procedure Combination (DPC) data is one of the standardized patient data formats in the Japanese medical field. DPC comprehensively collects information related to patient diagnoses and treatments and is used for improving the quality of medical care and economic analysis. DPC records information on diagnoses, surgeries, procedures, etc. in a specific coded format, which makes it easier to integrate data from different medical institutions and facilities. DPC data comprehensively records information on the medical procedures and medical expenses received by patients, including length of hospital stay, surgical procedures, test results, and medication prescriptions. Anonymized DPC information can be obtained, for example, from the Anonymous Medical Insurance Related Information Database (NDB) of the Ministry of Health, Labor, and Welfare of Japan.

[0034] An electronic medical record is a system that digitizes traditional paper medical records (medical history and treatment records) and records and manages medical information in electronic format, with the aim of improving the efficiency of information management and medical service provision in medical settings and improving the quality of medical care.Electronic medical records can centrally manage patient medical information, and electronic medical record data includes a variety of information such as consultation records, prescriptions, test results, and image data.

[0035] Health checkup data is information collected to assess an individual's health condition and disease risk. It is generally used by doctors and health managers to understand the health status of patients and examinees and to suggest necessary care and preventative measures. Health checkup data includes physical information (height, weight, blood pressure, etc.), lifestyle habits (smoking, drinking, exercise, etc.), blood test results, and imaging diagnostic results (X-ray, ultrasound, MRI, etc.).

[0036] Patient registry data is a database used to collect, manage, and analyze information on patients with specific diseases or symptoms. It is used for various purposes, including clinical research, healthcare quality improvement, and healthcare policy formulation. Patient registries may include basic patient demographic information (e.g., age, gender, race) and medical records (e.g., diagnoses, treatments, and test results). They may also collect tracking information on treatment effectiveness, side effects, and complications. Patient registry data can be collected in a variety of ways, including through information provided by patients and healthcare providers, through the use of electronic health records (EHRs) and medical claims data, or through data collection based on specific research protocols such as clinical trials. Patient registry data is generally anonymized so that individual patients cannot be identified.

[0037] Data from wearable devices can provide information about a patient's health status and activity level. These devices can collect data such as heart rate, sleep patterns, steps taken, and physical activity. This data can identify trends in a patient's health and lifestyle and help medical professionals make appropriate diagnoses and provide treatment. Examples of wearable devices include smart watches, fitness trackers, and smart rings.

[0038] In some embodiments, the target variable (also called the ground truth label) of the data used to train the machine learning model may be the presence or absence of Fabry disease in a previously diagnosed patient and / or the presence or absence of suspicion of Fabry disease.

[0039] In some embodiments, the machine learning model is trained to output a value indicative of the likelihood that the patient has Fabry disease. Training the machine learning model generally involves the following steps: 1. Data collection and preprocessing: First, collect data appropriate to the problem. This data will be used to train the model. After collecting the data, preprocessing is performed, such as removing unnecessary parts, handling missing values, and normalizing the data. 2. Data Split: Split the data into three datasets: training dataset, validation dataset, and test dataset. The training dataset is used to train the model, the validation dataset is used for parameter adjustment and hyperparameter tuning of the model, and the test dataset is used to evaluate the final model. 3. Model selection: Select an appropriate model depending on the nature of the problem. For example, logistic regression, random forest, neural network, etc. are commonly used for classification problems. 4. Model Building: The selected model is implemented and fitted to the training dataset. In this process, the model parameters are adjusted to fit the data. 5. Train the model: The model is trained on the training dataset. This process allows the model to learn patterns and relationships in the data. Training is done using an optimization algorithm such as gradient descent. 6. Model Evaluation: The model is evaluated on the validation dataset to evaluate its performance. Evaluation metrics include precision, recall, precision, F1 score, etc. 7. Model tuning: Adjust the model's hyperparameters and structure to improve performance on the validation dataset, which will hopefully result in the model making better predictions. 8. Final evaluation: The final model is evaluated on the test dataset to see how well it performs. This step determines how well the model generalizes to unseen data. 9. Deployment: Finally, the trained model is deployed in production and ready to make predictions and classifications on new data.

[0040] In some embodiments, the machine learning model may be selected from the group consisting of decision trees, random forests, LightGBM, stacking (e.g., LR, RF, LGBM stacking), logistic regression, lasso regression, support vector machines, multilayer perceptrons, neural networks, and combinations thereof.

[0041] A decision tree is a method for recognizing patterns by dividing data into groups along a tree structure using conditional branching. A decision tree learns patterns and relationships in data by dividing a set of data based on simple questions and their conditions. This allows it to solve classification and regression problems. A decision tree has a tree structure, starting from a root node, with each node dividing the data based on a specific question. Each node generates new nodes (child nodes) through each branch, eventually reaching a leaf node (end). The leaf node outputs the final classification or prediction. Decision trees are intuitively easy to understand, making the constructed model easy to interpret. Furthermore, the importance of each feature can be evaluated during the model construction process. Furthermore, they are applicable even when the data has nonlinear relationships.

[0042] Random forest is an ensemble learning method in machine learning. Ensemble learning refers to the combination of multiple models to make a final prediction. Random forests are constructed by combining multiple decision trees. Random forests construct multiple decision trees and combine their results to make a final prediction. Each decision tree is trained independently. In random forests, multiple datasets are generated by randomly sampling from the original dataset. This technique is called bootstrap sampling. Each decision tree is trained using a different bootstrap sample. Only randomly selected features are considered when splitting each decision tree. This prevents overfitting of individual decision trees and improves the generalization performance of the overall model. After all decision trees have made their predictions, random forests make a final prediction by using majority voting in classification or averaging in regression. Random forests exhibit high predictive performance with relatively little data preprocessing and are applied to a variety of problems. Furthermore, feature importance can be easily evaluated, making them known as highly interpretable models.

[0043] LightGBM is a gradient boosting framework. Gradient boosting builds weak predictive models (usually decision trees) in order, learning to correct the errors of the previous model. LightGBM is known as a particularly fast and efficient implementation of gradient boosting. LightGBM uses a histogram-based algorithm, which builds a histogram of gradients for each feature in the dataset and uses an approximation algorithm to find efficient splits. LightGBM can perform effective modeling even for large datasets and data with high-dimensional features.

[0044] Logistic regression is a machine learning model used for classification problems. It is primarily used for two-class classification, but can be extended to multi-class classification. Logistic regression predicts the probability that input data belongs to each class by applying a linear combination of input variables to a logistic function (or sigmoid function). The sigmoid function returns a value between 0 and 1, which can be interpreted as a probability. Logistic regression works well for linearly separable problems, but is not suitable for problems with nonlinear relationships. In such cases, other methods such as kernel-based support vector machines and neural networks are considered.

[0045] Lasso regression is a widely used technique in statistics and machine learning, and is particularly effective for data with a large number of predictor variables. Lasso, short for "Least Absolute Shrinkage and Selection Operator," is a type of linear regression that introduces a penalty term (regularization term) based on the sum of the absolute values ​​of the parameters. Lasso regression has the ability to remove unnecessary predictor variables (features) from the model and select only important variables. This is achieved by suppressing some coefficients completely to zero using the penalty term. When multiple predictor variables are highly correlated, problems can arise with standard linear regression, but lasso regression helps alleviate this. Furthermore, reducing the number of variables makes the model simpler, making it easier to interpret and analyze.

[0046] Support vector machines (SVMs) are one of the powerful algorithms for machine learning classification and regression. They are primarily used for classification problems, but can also be applied to regression problems. The basic idea of ​​SVMs is to find the optimal decision boundary (hyperplane) for classifying data. This hyperplane is a boundary in multidimensional space, such as a line or plane, that best separates the data. SVMs attempt to find a decision boundary with the largest margin (distance) between data of different classes, which improves generalization performance. SVMs can also be applied to nonlinear data sets using the kernel trick, which allows them to map data into a high-dimensional space and find nonlinear boundaries. SVMs have excellent generalization ability with a relatively low risk of overfitting.

[0047] A neural network is a type of machine learning model based on a mathematical model inspired by how the human brain works. A neural network consists of multiple layers, each of which can consist of multiple neurons (nodes). These neurons receive input from the previous layer, sum up each input with a weight, and apply an activation function to the sum to generate an output. A neural network typically consists of the following main layers: 1. Input layer: This layer receives data, and each input is associated with a neuron. 2. Hidden layer: Layers between the input and output layers that model complex relationships in the data. When there are one or more hidden layers, they are called "deep neural networks." 3. Output layer: The final layer of a neural network, which outputs the model's predictions or classifications.

[0048] Training a neural network typically involves the following steps: 1. Data preparation: Prepare the input data and the corresponding correct labels (training data). 2. Network definition: Define the neural network structure and select the appropriate number of layers and neurons. 3. Training the model: Using the input data, the weights and biases of the network are adjusted to move the output closer to the ground truth label. This is typically done using gradient descent or a derivative algorithm. 4. Model evaluation: The trained model is applied to test or validation data and its performance is evaluated using metrics such as classification accuracy and prediction precision.

[0049] Multilayer Perceptron (MLP) is a type of artificial neural network (ANN) characterized by the existence of multiple layers (hidden layers). MLP consists of an input layer, multiple hidden layers, and an output layer. Each layer consists of multiple neurons (nodes), which are connected to neurons in adjacent layers. MLP has the following structure and operation: 1. Input layer: The data given to the model is input to this layer. The number of nodes in the input layer is equal to the number of features. 2. Hidden Layer: There are one or more hidden layers, each containing multiple neurons. Each neuron is connected to all neurons in the previous layer, and each connection is assigned a weight. An activation function is applied to introduce nonlinearity. Common activation functions include sigmoid, ReLU, and tanh (hyperbolic tangent). 3. Output layer: This layer generates the results output by the model. The number of nodes in the output layer varies depending on the type of problem. For regression problems, there is one node, while for classification, there are as many nodes as there are classes. Common activation functions are linear for regression problems, sigmoid for binary classification problems, and softmax for multi-class classification problems. MLPs can model complex nonlinear relationships between inputs and outputs by learning appropriate weight combinations. The training process typically uses an algorithm called backpropagation, which updates the weights to minimize the error between the model's predictions and the target.

[0050] Stacking machine learning models is a technique for combining multiple different models to build a more powerful predictive model. For example, LR (Logistic Regression), RF (Random Forest), and LGBM (LightGBM) are machine learning models each with a different algorithm. Stacking these models, although not limited to these, can utilize the strengths of each model and compensate for each other's weaknesses. Any model, including the models disclosed herein, can be used for stacking. Stacking is generally performed as follows: 1. First stage: Using the training data, multiple base models are trained. For example, different types of models, such as LR, RF, and LGBM, are used, each of which tends to capture different features. 2. Second stage: The prediction results of each base model trained in the first stage are treated as new features. Based on the training data, another model called a meta-model is trained. This model takes the prediction results of the base models as input and makes the final prediction. 3. Prediction: First, predictions are made using each basic model in the first stage for the test data. Then, using the prediction results, the final prediction is made using the meta-model in the second stage.

[0051] In some embodiments, the predictive model is a machine learning model trained to output a value indicating the likelihood that a patient has Fabry disease in response to input of one or more explanatory variables, including a patient's diagnosis history, test history, prescription history, medical treatment history, and / or attributes. Here, the one or more explanatory variables may be, for example, one, two, or three or more. However, this number refers to the number of categories, assuming the classification of the patient's diagnosis history, test history, prescription history, medical treatment history, and attributes, and is not related to the number of data belonging to each category.

[0052] In some embodiments, a patient's diagnosis history, test history, prescription history, medical procedure history, and / or attributes may be extracted from real-world data such as those described above.

[0053] In some embodiments, the predictive model may be a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients diagnosed by a physician as having or suspected of having Fabry disease.

[0054] In some embodiments, the predictive model may be a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients who have been diagnosed by a physician as not having or not suspected of having Fabry disease.

[0055] In some embodiments, a diagnosis in the history of diagnoses includes a diagnosis that confirms the disease.

[0056] In some embodiments, a diagnosis in the diagnostic history includes a diagnosis indicating suspicion of disease.

[0057] In some embodiments, the diagnoses in the history of diagnoses are disseminated intravascular coagulation, type 2 diabetes, congestive heart failure, Sjogren's syndrome, rheumatoid arthritis, acute myocardial infarction, acute progressive glomerulonephritis, rapidly progressive glomerulonephritis, angina pectoris, blood clotting disorders, hypothyroidism, hyperthyroidism, valvular heart disease, heart failure, deep vein thrombosis, decreased renal function, systemic lupus erythematosus, iron deficiency anemia, iron deficiency anemia, diabetes, Includes the diagnosis of diseases selected from the group consisting of urinary tract infection, cerebral infarction, pneumonia, arrhythmia, collagen disease, mitral valve regurgitation, type 2 diabetes mellitus (no diabetic complications), congestive heart failure, suspected disease number, angina pectoris, hypothyroidism, hyperphosphatemia, heart failure, renal anemia, sleep apnea syndrome, amplified valve regurgitation, dehydration, hypertrophic cardiomyopathy, arrhythmia, peripheral neuropathy, cholinergic urticaria, plasma cell myeloma, chronic heart failure, and chronic renal failure. The diagnoses in the diagnostic history may be, for example, diagnoses of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, or 43 diseases selected from the group above.

[0058] In some embodiments, the diagnoses in the diagnostic history include (suspected) disseminated intravascular coagulation, (suspected) type 2 diabetes, (suspected) congestive heart failure, (suspected) Sjogren's syndrome, (suspected) rheumatoid arthritis, (suspected) acute myocardial infarction, (suspected) acute progressive glomerulonephritis, (suspected) rapidly progressive glomerulonephritis, (suspected) angina pectoris, (suspected) blood coagulation disorder, (suspected) hypothyroidism, (suspected) hyperthyroidism, (suspected) valvular heart disease, (suspected) heart failure, (suspected) deep vein thrombosis, (suspected) decreased renal function, (suspected) systemic lupus erythematosus, and (suspected) iron deficiency anemia. The diagnosis includes a diagnosis of a disease selected from the group consisting of blood, (suspected) iron deficiency anemia, (suspected) diabetes, (suspected) urinary tract infection, (suspected) cerebral infarction, (suspected) pneumonia, (suspected) arrhythmia, (suspected) collagen disease, mitral valve regurgitation, type 2 diabetes with no diabetic complications, congestive heart failure, number of suspected diseases, angina pectoris, hypothyroidism, hyperphosphatemia, heart failure, age at diagnosis, renal anemia, sleep apnea syndrome, amplified valvular regurgitation, dehydration, hypertrophic cardiomyopathy, arrhythmia, peripheral neuropathy, cholinergic urticaria, plasma cell myeloma, chronic heart failure, and chronic renal failure. Note that, here, (suspected) indicates a diagnosis that indicates that the disease is suspected. The diagnosis in the diagnosis history is, for example, a diagnosis of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, or 43 diseases selected from the above group.

[0059] In some embodiments, the testing history includes a number of tests. In some embodiments, the tests in the testing history include a creatinine test.

[0060] In some embodiments, the attributes include attributes selected from the group consisting of sex, age, age at diagnosis, and suspected disease number. The attributes can be, for example, one, two, or three attributes selected from the group consisting of these.

[0061] In some embodiments, the predictive model is a rule-based filtering model constructed to output a value indicative of the likelihood that a patient has Fabry disease. The likelihood value may be, for example, either likely or unlikely. The likelihood value may include, for example, one or more values ​​intermediate between likely and unlikely.

[0062] In some embodiments, the predictive model is a rule-based filtering model that is constructed to output a value indicating the likelihood that a patient has Fabry disease in response to input of one or more explanatory variables, including a patient's diagnosis history, test history, prescription history, medical treatment history, and / or attributes. Here, the one or more explanatory variables may be, for example, 1, 2, or 3, but this number refers to the number of categories, assuming the classification of the patient's diagnosis history, test history, prescription history, medical treatment history, and attributes, and is not related to the number of data belonging to each category.

[0063] In some embodiments, the predictive model is an unsupervised learning model that performs clustering based on one or more input data, including a patient's diagnosis history, test history, prescription history, medical treatment history, and / or attributes. Clustering can be performed, for example, by k-means or a method derived therefrom. Here, the one or more data can be, for example, one, two, or three pieces of data. However, this number refers to the number of categories, assuming the categories of a patient's diagnosis history, test history, prescription history, medical treatment history, and attributes, and is not related to the number of data belonging to each category.

[0064] In some embodiments, the predictive model may be based on the results of the clustering: a value indicating the likelihood that a patient has Fabry disease may be calculated, for example, by calculating the distance from the centroid of each cluster.

[0065] In some embodiments, a system according to the present disclosure provides a medical representative (MR) with information about patients and / or physicians in the MR's area of ​​responsibility. In some embodiments, the information is provided in response to a request from the MR (e.g., a search or other input). In some embodiments, the system determines certain conditions (e.g., the passage of a certain period of time or the addition of new information) and automatically provides the information to the MR. Such information can be useful in improving the efficiency of the MR's work.

[0066] In some embodiments, the system of the present disclosure may suggest to a medical representative which hospitals and / or doctors the medical representative should contact. For example, if there are patients with possible Fabry disease in the medical representative's area of ​​responsibility, information about hospitals and / or doctors the patients may be seeing may be presented. Such information may be useful in improving the efficiency of the medical representative's work.

[0067] In some embodiments, the hospital that the medical representative should contact is determined based on the number of patients who may be suffering from Fabry disease and who may be visiting the hospital. The number of patients may be, for example, 1 or more, 2 or more, 3 or more, 5 or more, 10 or more, 20 or more, or 30 or more. The hospital that should be contacted may also be determined based on a change (e.g., an increase) in the number of patients. For example, a hospital may be determined to be one that the medical representative should contact if the number of patients changes from 0 to 1, 2, or 3 or more, or from 1 to 2, 3, or 4 or more. In some embodiments, for example, based on the change in the number of patients, the system according to the present disclosure may send a notification or alert to a user terminal.

[0068] In some embodiments, the hospital that the medical representative should contact is determined based on the number of doctors affiliated with that hospital who have the potential to diagnose and / or treat Fabry disease. The number of doctors can be, for example, 1 or more, 2 or more, 3 or more, 5 or more, 10 or more, 20 or more, or 30 or more. The hospital that should be contacted may also be determined based on a change (e.g., an increase) in the number of doctors. For example, a hospital may be determined to be one that the medical representative should contact if the number of doctors changes from 0 to 1, 2, or 3 or more, or from 1 to 2, 3, or 4 or more. In some embodiments, for example, based on the change in the number of doctors, the system according to the present disclosure may send a notification or alert to a user terminal.

[0069] Information on doctors who may be able to diagnose and / or treat Fabry disease In some embodiments, the server of the present disclosure stores information about physicians who may be able to diagnose and / or treat Fabry disease. In some embodiments, the physician information includes a list of one or more hospitals to which the physician is affiliated. The hospital may be identified, for example, from information such as the physician's name. If the physician is affiliated with multiple facilities or institutions, a list of hospitals may be created. The hospitals included in the list may be ranked based on the type and number of medical departments, the number of staff, the number of hospital beds, etc.

[0070] In some embodiments, the physician information includes a value indicating the likelihood that the physician will be able to diagnose and / or treat Fabry disease. In some embodiments, the value indicating the likelihood that the physician will be able to diagnose and / or treat Fabry disease may be, for example, an integer or real number ranging from 0 to 100, but is not limited thereto. In some embodiments, the value indicating the likelihood that the physician will be able to diagnose and / or treat Fabry disease may have any threshold value; for example, only physicians having a value of 50 or greater, 60 or greater, 70 or greater, or 80 or greater may be considered to be physicians who can potentially diagnose and / or treat Fabry disease. In some embodiments, the value indicating the likelihood that the physician will be able to diagnose and / or treat Fabry disease may be stratified into, for example, two, three, four, five, six, seven, eight, nine, or ten levels. For example, in the case of a two-level value, the value indicating the likelihood that the physician will be able to diagnose and / or treat Fabry disease may be stratified into two levels: "low likelihood" and "high likelihood." For example, in the case of a 3-point scale, the value indicating the likelihood that a doctor will be able to diagnose and / or treat Fabry disease may be stratified into three categories: "high," "medium," or "low." For example, in the case of a 10-point scale, the value indicating the likelihood that a doctor will be able to diagnose and / or treat Fabry disease may be ranked using a number from 1 to 10.

[0071] In some embodiments, the value indicating the likelihood that a doctor can diagnose and / or treat Fabry disease is determined based on one or more items including the doctor's gender, age, hometown, alma mater, year of graduation, history of working facility, history of medical department, history of accessing digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers related to Fabry disease. Furthermore, in some embodiments, the value indicating the likelihood that a doctor can diagnose and / or treat Fabry disease is determined based on the number or frequency of accesses to digital content related to Fabry disease. The digital content may be, for example, a web page, video, or audio related to Fabry disease. The video and / or audio may be provided in any format, for example, live streaming, on-demand streaming, etc.

[0072] In some embodiments, the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease is a predicted value or a stratified predicted value obtained by a predictive model. In some embodiments, the predictive model may be, but is not limited to, a machine learning model, a statistical model, a simulation model, a rule-based model, a time series model, or a symbolic regression model.

[0073] In some embodiments, the predictive model may output a value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease, for example, but not limited to, a real number between 0 and 1, or an integer or real number between 0 and 100. In some embodiments, the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease may have an arbitrary threshold value, for example, only physicians having a value of 0.5 or more, 0.6 or more, 0.7 or more, or 0.8 or more, or a value of 50 or more, 60 or more, 70 or more, or 80 or more may be considered to be physicians who can potentially diagnose and / or treat Fabry disease, and other values ​​may be set to 0.

[0074] In some embodiments, the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease may be stratified into, for example, 2, 3, 4, 5, 6, 7, 8, 9, or 10 levels. For example, in the case of a 2-level scale, the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease may be stratified into two levels: "low likelihood" and "high likelihood." For example, in the case of a 3-level scale, the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease may be stratified into three levels: "high," "medium," or "low." For example, in the case of a 10-level scale, the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease may be ranked using a number from 1 to 10.

[0075] In some embodiments, the prediction model is a machine learning model trained to output a value indicating the likelihood that a doctor will be able to diagnose and / or treat Fabry disease in response to input of one or more explanatory variables, including the doctor's gender, age, hometown, alma mater, year of graduation, work facility history, medical department history, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers about Fabry disease. Here, the one or more explanatory variables may be, for example, one, two, or three. However, this number refers to the number of categories, assuming the categories are the doctor's gender, age, hometown, alma mater, year of graduation, work facility history, medical department history, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and history of writing papers about Fabry disease, and is not related to the number of data belonging to each category.

[0076] In some embodiments, the behavioral history of access to digital content related to Fabry disease may include the number of views of videos related to Fabry disease, the number of unique actions, access trends by channel, the average or variance of digital access frequency, the average or variance of the number of activities on the day of digital access behavior, etc.

[0077] In some embodiments, the Fabry disease publication history may include co-author information.

[0078] In some embodiments, the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers related to Fabry disease of physicians known to be able to diagnose and / or treat Fabry disease.

[0079] In some embodiments, the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers related to Fabry disease of physicians known to be unable to diagnose and / or treat Fabry disease.

[0080] In some embodiments, the target variable (also called the correct label) of the data used to train the machine learning model may be whether or not a doctor has a case of Fabry disease.

[0081] In some embodiments, the machine learning model is selected from the group consisting of decision trees, random forests, LightGBM, stacking (e.g., LR, RF, LGBM stacking), logistic regression, lasso regression, support vector machines, multilayer perceptrons, neural networks, and combinations thereof.

[0082] In some embodiments, the prediction model is a rule-based filtering model constructed to output a value indicating the likelihood that a doctor will be able to diagnose and / or treat Fabry disease in response to input of one or more explanatory variables including the doctor's gender, age, hometown, alma mater, year of graduation, work facility history, medical department history, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers about Fabry disease. Here, the one or more explanatory variables may be, for example, one, two, or three, and this number refers to the number of categories, assuming the categories are the doctor's gender, age, hometown, alma mater, year of graduation, work facility history, medical department history, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and history of writing papers about Fabry disease, and is not related to the number of data belonging to each category.

[0083] In some embodiments, the prediction model is an unsupervised learning model that performs clustering based on one or more input data, including the doctor's gender, age, hometown, alma mater, year of graduation, work facility history, medical department history, history of access to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers related to Fabry disease. Clustering can be performed, for example, by the k-means algorithm or a method derived therefrom. Here, the one or more data items can be, for example, one, two, or three. However, this number refers to the number of categories, assuming that the categories are the doctor's gender, age, hometown, alma mater, year of graduation, work facility history, medical department history, history of access to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and history of writing papers related to Fabry disease, and is not related to the number of data items belonging to each category.

[0084] In some embodiments, a system according to the present disclosure provides a medical representative with information about patients and / or physicians in the medical representative's area of ​​responsibility. In some embodiments, the information is provided in response to a request from the medical representative (e.g., a search or other input). In some embodiments, the system determines certain conditions (e.g., the passage of a certain period of time or the addition of new information) and automatically provides the information to the medical representative. Such information can be useful in improving the efficiency of the medical representative's work.

[0085] In some embodiments, the system of the present disclosure may suggest to the medical representative which hospitals and / or doctors the medical representative should contact. For example, if there is a doctor in the medical representative's area who may be able to diagnose and / or treat Fabry disease, information about the hospital or institution to which the doctor belongs may be presented. Such information may be useful in improving the efficiency of the medical representative's work.

[0086] How to predict whether a patient has Fabry disease In one aspect, the present disclosure relates to a method for predicting the likelihood that a patient has Fabry disease. In some embodiments, the prediction result obtained by the method according to the present disclosure can be used to assist medical representatives in their activities. In another embodiment, the prediction result obtained by the method according to the present disclosure can be used to assist physicians in diagnosing Fabry disease.

[0087] In some embodiments, the method of predicting the likelihood that a patient has Fabry disease of the present disclosure includes inputting at least a portion of the patient's real-world data, including medical insurance claim data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from a wearable device, into a prediction model, and obtaining an output from the prediction model indicating the likelihood that the patient has Fabry disease. In some embodiments, the method of the present disclosure can be executed on any computer.

[0088] In some embodiments, the value indicating the likelihood that a patient has Fabry disease is a predicted value or a stratified predicted value from a prediction model.

[0089] In some embodiments, the predictive model is a machine learning model trained to take as input at least a portion of a patient's real-world data, including medical receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from wearable devices, and output a value indicating the likelihood that the patient has Fabry disease.

[0090] In some embodiments, the predictive model is a machine learning model trained to output a value indicating the likelihood that a patient has Fabry disease in response to input of one or more explanatory variables, including a patient's diagnostic history, testing history, prescription history, medical treatment history, and / or attributes.

[0091] In some embodiments, the predictive model is a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients diagnosed by a physician as having or suspected of having Fabry disease.

[0092] In some embodiments, the predictive model is a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients who have been diagnosed by a physician as not having or not suspected of having Fabry disease.

[0093] In some embodiments, a diagnosis in the history of diagnoses includes a diagnosis that confirms the disease.

[0094] In some embodiments, a diagnosis in the diagnostic history includes a diagnosis indicating suspicion of disease.

[0095] In some embodiments, the diagnoses in the history of diagnoses are disseminated intravascular coagulation, type 2 diabetes, congestive heart failure, Sjogren's syndrome, rheumatoid arthritis, acute myocardial infarction, acute progressive glomerulonephritis, rapidly progressive glomerulonephritis, angina pectoris, blood clotting disorders, hypothyroidism, hyperthyroidism, valvular heart disease, heart failure, deep vein thrombosis, decreased renal function, systemic lupus erythematosus, iron deficiency anemia, iron deficiency anemia, diabetes, Includes the diagnosis of diseases selected from the group consisting of urinary tract infection, cerebral infarction, pneumonia, arrhythmia, collagen disease, mitral valve regurgitation, type 2 diabetes mellitus (no diabetic complications), congestive heart failure, suspected disease number, angina pectoris, hypothyroidism, hyperphosphatemia, heart failure, renal anemia, sleep apnea syndrome, amplified valve regurgitation, dehydration, hypertrophic cardiomyopathy, arrhythmia, peripheral neuropathy, cholinergic urticaria, plasma cell myeloma, chronic heart failure, and chronic renal failure.

[0096] In some embodiments, the testing history includes a number of tests. In some embodiments, the tests in the testing history include a creatinine test.

[0097] In some embodiments, the attributes include attributes selected from the group consisting of sex, age, age at diagnosis, and suspected disease number.

[0098] In some embodiments, the machine learning model is selected from the group consisting of decision trees, random forests, LightGBM, stacking (e.g., LR, RF, LGBM stacking), logistic regression, lasso regression, support vector machines, multilayer perceptrons, neural networks, and combinations thereof.

[0099] In one aspect, the present disclosure relates to a method for identifying hospitals where patients with Fabry disease may be treated. In some embodiments, the method for identifying hospitals where patients with Fabry disease may be treated includes the steps of identifying patients who are predicted to have a possibility of having Fabry disease by a method for predicting the possibility that a patient has Fabry disease according to the present disclosure, identifying an area where a hospital where the patient is treated is located based on attribute information of the identified patient, and identifying hospitals located in the identified area.

[0100] A computer program that predicts a patient's likelihood of having Fabry disease In one aspect, the present disclosure relates to a computer program for predicting the likelihood that a patient has Fabry disease. In some embodiments, the computer program for predicting the likelihood that a patient has Fabry disease of the present disclosure, when implemented on a computer, includes instructions for causing a processor to execute the following steps: inputting at least a portion of the patient's real-world data, including medical receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data derived from a wearable device, into a prediction model; and obtaining, as an output from the prediction model, a value indicating the likelihood that the patient has Fabry disease.

[0101] In some embodiments, the computer program according to the present disclosure may be a computer program for causing a computer to perform a method according to the present disclosure for predicting the likelihood that a patient has Fabry disease.

[0102] In some embodiments, a computer program according to the present disclosure may be stored in a computer-readable recording medium. Thus, in one aspect, the present disclosure relates to a computer-readable recording medium having a computer program according to the present disclosure recorded thereon. Examples of computer-readable recording media include, but are not limited to, hard disk drives (HDDs), solid-state drives (SSDs), USB flash drives, CD-ROMs, DVD-ROMs, Blu-ray (registered trademark) discs, memory cards, and magnetic tapes.

[0103] In one aspect, the present disclosure relates to a computer-readable medium having non-transiently stored thereon instructions that, when executed by a processor, perform the following steps: (i) inputting a patient's real-world data into a predictive model (S100); (ii) predicting the likelihood that the patient has Fabry disease using the predictive model (S110); and (iii) outputting a value indicative of the likelihood that the patient has Fabry disease (S120).

[0104] Furthermore, in one aspect, the present disclosure relates to a trained machine learning model that can be used to cause a computer to perform the method of predicting the likelihood that a patient has Fabry disease according to the present disclosure. The trained machine learning model can be accessed, for example, via an API that receives an external request, passes it to the model, and returns a prediction result.

[0105] Method for generating a trained machine learning model for use in predicting the likelihood that a patient has Fabry disease In one aspect, the present disclosure relates to a method for generating a trained machine learning model for use in predicting the likelihood that a patient has Fabry disease. In some embodiments, the model generation method of the present disclosure uses training data in which at least a portion of the patient's real-world data, including medical receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from a wearable device, is used as an explanatory variable, and the presence or absence of Fabry disease or the presence or absence of suspicion of Fabry disease in the patient is used as a response variable.

[0106] In some embodiments, at least a portion of the patient's real-world data includes the patient's diagnosis history, test history, prescription history, medical procedure history, and / or attributes.

[0107] Device for predicting the likelihood that a patient has Fabry disease In one aspect, the present disclosure relates to a device for predicting the likelihood that a patient has Fabry disease. In some embodiments, the device for predicting the likelihood that a patient has Fabry disease includes an input unit for inputting at least a portion of the patient's real-world data, including medical receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from a wearable device, into a prediction model; a prediction unit for predicting the likelihood that the patient has Fabry disease using the prediction model; and an output unit for outputting a value indicating the predicted likelihood that the patient has Fabry disease. In some embodiments, the device for predicting the likelihood that a patient has Fabry disease may be part of a system for supporting medical representatives (MRs).

[0108] In some embodiments, a device according to the present disclosure comprises means for performing a method according to the present disclosure for predicting the likelihood that a patient has Fabry disease.

[0109] In some embodiments, a device according to the present disclosure is a device that includes a processor capable of executing instructions of a computer program stored therein that predicts the likelihood that a patient according to the present disclosure has Fabry disease.

[0110] In some embodiments, the device according to the present disclosure is a device for predicting the likelihood that a patient has Fabry disease, including a prediction unit that makes a prediction using a trained machine learning model according to the present disclosure for use in predicting the likelihood that a patient has Fabry disease.

[0111] In some embodiments, the device according to the present disclosure includes a means for inputting data, such as a keyboard, a mouse, or the like.

[0112] In some embodiments, the device according to the present disclosure includes a central processing unit (CPU) connected to a keyboard, mouse, etc. for inputting data, connected to a hard disk, flash memory, etc. as a storage unit, and connected to memory (storage means) such as ROM, RAM, etc.

[0113] Examples of means for outputting data including prediction results include monitors, printers, etc. Other examples of output means include means for storing data in storage means such as hard disks, flash memories, ROMs, RAMs, etc.

[0114] The apparatus according to the present disclosure may include a means for storing a computer program for predicting the likelihood that a patient according to the present disclosure has Fabry disease. Examples of the means for storing the program include a hard disk, a flash memory, and the like. Such a storage means may be connected via a communication line. In other words, the apparatus according to the present disclosure may be part of a system obtained by connecting via a communication line to a device including the means for storing the program.

[0115] Fig. 1 is a schematic diagram showing an exemplary embodiment of an apparatus according to the present disclosure. In Fig. 1, reference numeral 100 denotes a computer, which includes a control unit 101, a storage unit 102, a peripheral device I / F unit 103, an input unit 104, a display unit 105, and a communication unit 106, all of which are connected by a bus 110. Note that this configuration is merely an example, and various other configurations may be adopted as appropriate.

[0116] The control unit 101 is composed of a CPU (Central Processing Unit), ROM (Read Only Memory), RAM (Random Access Memory), etc. The CPU loads programs stored in the storage unit 102, ROM, recording medium, etc. into a work memory area on the RAM and executes them, driving and controlling each device connected via the bus 110 and realizing the processing performed by the computer. The ROM is a non-volatile memory and stores programs such as the boot program and BIOS of the computer 100, data, etc. The RAM is a volatile memory and temporarily stores programs, data, etc. loaded from the storage unit 102, ROM, recording medium, etc., and also has a work area used by the control unit 101 when performing various processing. The storage unit 102 is, for example, an HDD (Hard Disk Drive), and stores the programs executed by the control unit 101 and various other data.

[0117] The peripheral device I / F (interface) unit 103 is a port for connecting the computer 100 to peripheral devices. The peripheral device I / F unit 103 is configured with a USB, IEEE1394, RS-232C, or the like. The connection with the peripheral devices may be wired or wireless. The input unit 104 has input devices such as a keyboard, a pointing device such as a mouse, and a numeric keypad, and issues operation instructions, operational instructions, data input, and the like to the computer 100. The display unit 105 is a logic circuit or device driver for displaying videos, images, and the like on a display device such as a liquid crystal panel. The input unit 104 and the display unit 105 can also be configured integrally as a touch display.

[0118] The communication unit 106 has a communication control device, a communication port, etc., and is a wired or wireless communication interface that mediates communication with the network 120. The bus 110 is a communication path that mediates the exchange of control signals, data signals, etc. between the devices. The network 120 can further be connected to an external server 130 and a database (or net storage) 140.

[0119] Methods for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease In one aspect, the present disclosure relates to a method for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease. In some embodiments, the method for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease comprises the steps of inputting data including the doctor's gender, age, hometown, alma mater, year of graduation, work facility history, medical department history, access behavior history of digital content related to Fabry disease, purchase history of medication related to Fabry disease, and / or history of writing papers on Fabry disease into a prediction model, and obtaining, as output from the prediction model, a value indicating the likelihood that the doctor will be able to diagnose and / or treat Fabry disease. In some embodiments, the prediction model makes a prediction based on the number or frequency of accesses to digital content related to Fabry disease.

[0120] In some embodiments, the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease is a predicted value or a stratified predicted value from a prediction model.

[0121] In some embodiments, the predictive model is a machine learning model trained to output a value indicating the likelihood that a doctor will be able to diagnose and / or treat Fabry disease in response to input of one or more explanatory variables including the doctor's gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers related to Fabry disease.

[0122] In some embodiments, the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers related to Fabry disease of physicians known to be able to diagnose and / or treat Fabry disease.

[0123] In some embodiments, the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers related to Fabry disease of physicians known to be unable to diagnose and / or treat Fabry disease.

[0124] In some embodiments, the machine learning model is selected from the group consisting of decision trees, random forests, LightGBM, stacking (e.g., LR, RF, LGBM stacking), logistic regression, lasso regression, support vector machines, multilayer perceptrons, neural networks, and combinations thereof.

[0125] In some embodiments, the predictive model is a rule-based filtering model constructed to output a value indicative of the likelihood that a physician will be able to diagnose and / or treat Fabry disease. The likelihood value may be, for example, either likely or unlikely. The likelihood value may include, for example, one or more values ​​intermediate between likely and unlikely.

[0126] In some embodiments, the predictive model may be based on the results of the clustering: a value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease may be calculated, for example, by calculating the distance from the centroid of each cluster.

[0127] In some embodiments, the method according to the present disclosure further comprises providing digital content related to Fabry disease to a physician and obtaining a history of access behavior to the digital content related to Fabry disease. In some embodiments, the physician can access the server providing the digital content related to Fabry disease at any time to view the content, and the history of access behavior is recorded on the server.

[0128] In one aspect, the present disclosure also relates to a method for identifying a hospital where a doctor who can diagnose and / or treat Fabry disease may be employed. In some embodiments, the method for identifying a hospital where a doctor who can diagnose and / or treat Fabry disease may be employed comprises the steps of identifying a doctor who is predicted to have the potential to diagnose and / or treat Fabry disease by a method for predicting the potential of a doctor being able to diagnose and / or treat Fabry disease according to the present disclosure, and identifying the hospital to which the doctor is employed from attribute information of the identified doctor.

[0129] A computer program that predicts the likelihood that a physician will be able to diagnose and / or treat Fabry disease In one aspect, the present disclosure relates to a computer program for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease. In some embodiments, the computer program for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease, when implemented on a computer, includes instructions for causing a processor to execute the following steps: inputting data including the doctor's gender, age, hometown, alma mater, year of graduation, work facility history, medical department history, access behavior history of digital content related to Fabry disease, purchase history of medications related to Fabry disease, and / or history of writing papers related to Fabry disease into a prediction model; and obtaining, as output from the prediction model, a value indicating the likelihood that the doctor will be able to diagnose and / or treat Fabry disease. In some embodiments, the prediction model makes a prediction based on the number or frequency of accesses to digital content related to Fabry disease.

[0130] In some embodiments, the computer program for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease may be a computer program for causing a computer to perform a method for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease according to the present disclosure.

[0131] In one aspect, the present disclosure also relates to a computer-readable recording medium having recorded thereon a computer program for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease.

[0132] In one aspect, the present disclosure relates to a computer-readable medium having non-transiently stored thereon instructions that, when executed by a processor, perform the following steps: (i) inputting attribute data of a doctor to be predicted into a prediction model (S200); (ii) predicting the likelihood that the doctor will be able to diagnose / treat Fabry disease using the prediction model (S210); and (iii) outputting a value indicating the likelihood that the doctor will be able to diagnose / treat Fabry disease (S220).

[0133] Furthermore, in one aspect, the present disclosure also relates to a trained machine learning model used to cause a computer to execute a method for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease. In some embodiments, the prediction model may be a machine learning model trained to output a value indicating the likelihood that a doctor will be able to diagnose and / or treat Fabry disease in response to input of one or more explanatory variables including the doctor's gender, age, hometown, alma mater, year of graduation, history of working facility, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers on Fabry disease.

[0134] Methods for generating trained machine learning models for use in predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease In one aspect, the present disclosure relates to a method for generating a trained machine learning model for use in predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease. In some embodiments, the method for generating a trained machine learning model for use in predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease uses training data in which the doctor's gender, age, hometown, alma mater, year of university graduation, work facility history, medical department history, access behavior history of digital content related to Fabry disease, purchase history of medication related to Fabry disease, and / or paper writing history on Fabry disease are used as explanatory variables, and whether the doctor will be able to diagnose and / or treat Fabry disease is used as a response variable.

[0135] A device for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease In one aspect, the present disclosure relates to an apparatus for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease. In some embodiments, the apparatus for predicting the likelihood that a doctor will be able to diagnose and / or treat Fabry disease includes an input unit for inputting data including the doctor's gender, age, hometown, alma mater, year of graduation, work facility history, medical department history, access behavior history of digital content related to Fabry disease, purchase history of medication related to Fabry disease, and / or history of writing papers on Fabry disease into a prediction model; a prediction unit for predicting the likelihood that the doctor will be able to diagnose and / or treat Fabry disease using the prediction model; and an output unit for outputting a value indicating the predicted likelihood that the doctor will be able to diagnose and / or treat Fabry disease. In some embodiments, the prediction model makes a prediction based on the number or frequency of accesses to digital content related to Fabry disease.

[0136] In some embodiments, a device according to the present disclosure comprises means for performing a method according to the present disclosure for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease.

[0137] In some embodiments, a device according to the present disclosure includes a processor capable of executing instructions of a computer program stored on a memory device that predicts the likelihood that a physician will be able to diagnose and / or treat Fabry disease.

[0138] In some embodiments, the device according to the present disclosure may be a device for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease, including a prediction unit that makes the prediction using a trained machine learning model that is trained to predict the likelihood that a physician will be able to diagnose and / or treat Fabry disease.

[0139] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Although any methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, some potentially preferred methods and materials are now described. All publications mentioned herein are incorporated by reference to disclose and describe the methods and / or materials in connection with which the publications are cited. It is understood that the present disclosure supersedes the disclosure of the incorporated publication in the case of a conflict.

[0140] Where a range of values ​​is described, unless the context clearly dictates otherwise, it is understood that each intervening value, to the tenth of the unit of the lower limit, between the upper and lower limits of that range is also specifically disclosed. Each smaller range between any stated or intervening value in a stated range and any other stated or intervening value within that stated range is also encompassed within the disclosure. The upper and lower limits of these smaller ranges may independently be included or excluded, and each range including either, either, or both limits in the smaller ranges is also encompassed within the invention, but the specifically excluded limit in the stated range is reserved. When a stated range includes one or both of the limits, ranges excluding either or both of those included limits are also included within the invention. The term "about" with respect to a numerical value means within 5%.

[0141] The embodiments described herein are intended to be merely exemplary, and those skilled in the art will be able to make numerous variations and modifications without departing from the spirit of the present invention. Furthermore, certain variations and modifications may produce less than optimal results, but still provide satisfactory results. All such variations and modifications are intended to be within the scope of the present invention as defined by the appended claims. Furthermore, any combination of the components disclosed herein, or any transformation of the disclosed expression into a method, apparatus, system, computer program, data structure, recording medium, or the like, is also valid as an aspect of the present disclosure. Therefore, details described regarding the method of the present disclosure may also be applied to the system, computer program, data structure, recording medium, or the like.

[0142] The present disclosure will be further understood by reference to the following examples. These examples are provided solely to illustrate the claimed disclosure; the disclosure is not limited in scope by the exemplified embodiments, which are intended only as illustrations of single aspects of the disclosure. Any functionally equivalent methods are within the scope of the present disclosure. Various modifications of the present disclosure, in addition to those described herein, will be apparent to those skilled in the art from the foregoing description. Such modifications are intended to fall within the scope of the appended claims. [Example]

[0143] Example 1: Predicting the likelihood that a patient has Fabry disease A Fabry disease patient prediction model was created based on patient data, including information on patient attributes, diagnostic history, prescription history, and medical procedures, including information on conditions other than Fabry disease. Explanatory variables can include patient attributes (age, sex), diagnostic history (number of diagnoses (number of confirmed diagnoses, number of suspected diagnoses), number of suspected diagnoses), and testing history (number of creatinine tests). The objective variables included diagnosis / prescription history of either Fabry disease prescription, Fabry disease diagnosis, or suspected Fabry disease.

[0144] The machine learning models used were logistic regression (LR), random forest (RF), LightGBM (LGBM), and stacking (LR, RF, LGBM stacking). Of the trained machine learning models generated, the AUCs of RF, LightGBM, and LR are shown in Figures 7, 9, and 11.

[0145] AUC is a common metric for evaluating the performance of machine learning models. The ROC curve plots the relationship between the true positive rate (TPR) and the false positive rate (FPR), and AUC represents the area under this curve. AUC is used to summarize model performance for binary classification problems. The true positive rate indicates the proportion of correct positive samples correctly classified by the model, while the false positive rate indicates the proportion of samples that are actually negative but are incorrectly classified as positive by the model. Generally, an AUC closer to 1 indicates better model performance, while an AUC closer to 0.5 indicates the model is making random classifications. Generally, an AUC below 0.5 indicates the model is completely random, and within the range of 0.5 to 1, a higher AUC is considered better performance.

[0146] As shown in Figures 7, 9, and 11, the AUC for RF was 0.75, the AUC for LGBM was 0.73, and the AUC for LR was 0.71. These results indicate that the generated machine learning model can make predictions with a certain degree of accuracy.

[0147] The system disclosed herein can also identify highly important features through analysis of trained machine learning models. The results of analyzing the importance of features in LR, RF, and LightGBM are shown in Figures 8, 10, and 12, respectively. In all models, age and the number of suspected diseases are highly important. Iron deficiency anemia, hypertension, and not hypertension are also considered to be highly important.

[0148] In this way, machine learning can be used to predict whether a patient has Fabry disease, and the prediction results can be useful for identifying facilities where undiagnosed patients who may have Fabry disease are visiting, and for medical representatives to carry out disease awareness activities.

[0149] Example 2: Predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease Based on data describing physician characteristics, we created a predictive model for physicians with the potential to diagnose and / or treat Fabry disease. Explanatory variables included physician attributes, affiliated institutions, publication information, and digital behavior information. The dependent variable included whether or not the physician had a history of Fabry disease.

[0150] We developed a model using a bagging method that combines multiple models based on logistic regression. The performance of the generated model on the validation data was 80% accuracy, 80% precision, 80% recall, 80% FI, and 80% ROC-AUC, confirming a certain level of accuracy in assessing the likelihood that a physician will be able to diagnose and / or treat Fabry disease. Furthermore, the results of the machine learning model analysis suggested that the physician's access history to digital content related to Fabry disease (number of digital accesses related to Fabry disease) and the physician's purchase history of medications related to Fabry disease (whether or not the physician's previous work facility had received medications for Fabry disease) were important features.

[0151] When this model was applied to 219,954 doctors who were considered to have no cases, 43,170 were predicted to have a "possible" diagnosis. This prediction result could be useful in situations where, for example, hospitals where patients with possible Fabry disease are visiting have already been identified to some extent, and MRs would like to provide information to doctors who have the potential to diagnose and / or treat Fabry disease. [Explanation of symbols]

[0152] 100...Computer 101 Control unit 102...Storage section 103 Peripheral device I / F section 104 Input section 105...Display section 106···Communications Department 110 Bus 120···Network 130...External Server 140···Database 210···Server 212 Control unit 214...Storage section 216···Communications Department 220 User terminal 222 Control unit 224... Output section 226···Communications Department 230···Network 240 Patient prediction device 250···Physician prediction device 300 Patient prediction device 310 Input section 320···Prediction section 330 Output section 400···Physician prediction device 410 Input section 420···Prediction section 430 Output section

Claims

1. A system for supporting medical representatives (MRs), A server having a communication unit for connecting to a user terminal and a storage unit storing information on patients and / or doctors; A user terminal having a communication unit for connecting to a server and an output unit for providing information to a user. and stored on the server a) information on patients who may be affected by Fabry disease; and / or b) A system that provides medical representatives with information on doctors who may be able to diagnose and / or treat Fabry disease via output to a user terminal.

2. Stored on the server a) information on patients who may have Fabry disease; and The system of claim 1, wherein b) information on doctors who have the potential to diagnose and / or treat Fabry disease is provided to medical representatives via output to a user terminal.

3. The system of claim 1 , wherein the patient information includes information about the area where the hospital where the patient may be treated is located.

4. 10. The system of claim 1, wherein the patient information includes a number of patients with a likelihood of having Fabry disease associated with a particular region.

5. 2. The system of claim 1, wherein information about areas with a high number of patients who may have Fabry disease is provided to medical representatives in preference to areas with a low number of patients.

6. The system of claim 1 , wherein the patient information includes a list of one or more hospitals where the patient may be seen.

7. 2. The system of claim 1, which provides medical representatives with information about doctors who belong to hospitals where patients who may have Fabry disease may visit and who may be able to diagnose and / or treat Fabry disease.

8. 10. The system of claim 1, wherein the patient information includes a value indicating the likelihood that the patient has Fabry disease.

9. The system of claim 8, wherein the value indicating the possibility that a patient has Fabry disease is a predicted value by a prediction model or a stratified value of the predicted value.

10. The system of claim 9, wherein the predictive model is a machine learning model trained to input at least a portion of the patient's real-world data, including medical receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from a wearable device, and output a value indicating the likelihood that the patient has Fabry disease.

11. The system of claim 9, wherein the predictive model is a machine learning model trained to output a value indicating the likelihood that a patient has Fabry disease in response to input of one or more explanatory variables including a patient's diagnosis history, test history, prescription history, medical treatment history, and / or attributes.

12. The system of claim 11, wherein the predictive model is a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients diagnosed by a physician as having or suspected of having Fabry disease.

13. The system of claim 11, wherein the predictive model is a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients who have been diagnosed by a physician as not having or not suspected of having Fabry disease.

14. The system of claim 11 , wherein the diagnoses in the history of diagnoses include diagnoses that confirm the disease.

15. The system of claim 11 , wherein a diagnosis in the diagnosis history includes a diagnosis indicating suspicion of disease.

16. The diagnosis in the diagnostic history was disseminated intravascular coagulation, type 2 diabetes, congestive heart failure, Sjögren's syndrome, rheumatoid arthritis, acute myocardial infarction, acute progressive glomerulonephritis, rapidly progressive glomerulonephritis, angina pectoris, blood coagulation disorder, hypothyroidism, hyperthyroidism, valvular heart disease, heart failure, deep vein thrombosis, decreased renal function, systemic lupus erythematosus, iron deficiency anemia, diabetes, urinary tract infection, cerebral infarction, pneumonia, and urinary tract infection. The system of claim 11, comprising the diagnosis of a disease selected from the group consisting of arrhythmia, collagen disease, mitral valve regurgitation, type 2 diabetes with or without diabetic complications, congestive heart failure, suspected disease, angina pectoris, hypothyroidism, hyperphosphatemia, heart failure, renal anemia, sleep apnea syndrome, amplified valvular regurgitation, dehydration, hypertrophic cardiomyopathy, arrhythmia, peripheral neuropathy, cholinergic urticaria, plasma cell myeloma, chronic heart failure, and chronic renal failure.

17. The system of claim 11 , wherein the tests in the test history include a creatinine test.

18. The system of claim 11 , wherein the attributes include attributes selected from the group consisting of gender, age, age at diagnosis, and number of suspected diseases.

19. 12. The system of claim 11, wherein the machine learning model is selected from the group consisting of a decision tree, a random forest, a LightGBM, stacking, logistic regression, lasso regression, a support vector machine, a multi-layer perceptron, a neural network, and combinations thereof.

20. The system of claim 1 , wherein the physician information includes a list of one or more hospitals to which the physician is affiliated.

21. 2. The system of claim 1, wherein the physician information includes a value indicating the likelihood that the physician will be able to diagnose and / or treat Fabry disease.

22. The system of claim 21, wherein the value indicating the likelihood that a doctor can diagnose and / or treat Fabry disease is determined based on one or more items including the doctor's gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers on Fabry disease.

23. 23. The system of claim 22, wherein the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease is determined based on the number or frequency of accesses to digital content related to Fabry disease.

24. The system of claim 21, wherein the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease is a predicted value by a prediction model or a stratified value of the predicted value.

25. The system of claim 24, wherein the predictive model is a machine learning model trained to output a value indicating the likelihood that a doctor will be able to diagnose and / or treat Fabry disease in response to input of one or more explanatory variables including the doctor's gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers on Fabry disease.

26. The system of claim 25, wherein the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, access behavior history of digital content related to Fabry disease, purchase history of medications related to Fabry disease, and / or history of writing papers on Fabry disease of physicians known to be able to diagnose and / or treat Fabry disease.

27. The system of claim 25, wherein the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department affiliation, access behavior history of digital content related to Fabry disease, purchase history of medications related to Fabry disease, and / or history of writing papers on Fabry disease of doctors known to be unable to diagnose and / or treat Fabry disease.

28. 26. The system of claim 25, wherein the machine learning model is selected from the group consisting of a decision tree, a random forest, a LightGBM, stacking, a logistic regression, a lasso regression, a support vector machine, a multi-layer perceptron, and a neural network, and combinations thereof.

29. 29. A system according to any one of claims 1 to 28, providing a medical representative with information on patients and / or doctors in the medical representative's area of ​​responsibility.

30. 29. A system according to any one of claims 1 to 28, which suggests to a medical representative which hospitals and / or doctors the medical representative should contact.

31. 31. The system of claim 30, wherein the hospital that the medical representative should contact is determined based on the number of patients who may be suffering from Fabry disease and who may be visiting the hospital.

32. 31. The system of claim 30, wherein the hospital that the medical representative should contact is determined based on the number of doctors affiliated with that hospital who have the potential to diagnose and / or treat Fabry disease.

33. 1. A method for predicting the likelihood that a patient has Fabry disease, comprising: Inputting at least a portion of the patient's real-world data, including receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from a wearable device, into a predictive model; and obtaining an output from the predictive model indicating the likelihood that the patient has Fabry disease. A method comprising:

34. The method of claim 33, wherein the value indicating the possibility that a patient has Fabry disease is a predicted value by a prediction model or a stratified predicted value.

35. The method of claim 33, wherein the predictive model is a machine learning model trained to input at least a portion of the patient's real-world data, including medical receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from a wearable device, and output a value indicating the likelihood that the patient has Fabry disease.

36. The method of claim 33, wherein the predictive model is a machine learning model trained to output a value indicating the likelihood that a patient has Fabry disease in response to input of one or more explanatory variables including a patient's diagnosis history, test history, prescription history, medical treatment history, and / or attributes.

37. The method of claim 36, wherein the predictive model is a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients diagnosed by a physician as having or suspected of having Fabry disease.

38. The method of claim 36, wherein the predictive model is a machine learning model trained using one or more explanatory variables including diagnostic history, testing history, prescription history, medical treatment history, and / or attributes of one or more patients who have been diagnosed by a physician as not having or not suspected of having Fabry disease.

39. 37. The method of claim 36, wherein a diagnosis in the history of diagnoses includes a diagnosis that confirms the disease.

40. 37. The method of claim 36, wherein a diagnosis in the diagnostic history includes a diagnosis indicating suspicion of disease.

41. The diagnosis in the diagnostic history was disseminated intravascular coagulation, type 2 diabetes, congestive heart failure, Sjögren's syndrome, rheumatoid arthritis, acute myocardial infarction, acute progressive glomerulonephritis, rapidly progressive glomerulonephritis, angina pectoris, blood coagulation disorder, hypothyroidism, hyperthyroidism, valvular heart disease, heart failure, deep vein thrombosis, decreased renal function, systemic lupus erythematosus, iron deficiency anemia, iron deficiency anemia, diabetes, urinary tract infection, cerebral infarction, pneumonia, The method of claim 36, comprising diagnosing a disease selected from the group consisting of arrhythmia, collagen disease, mitral valve regurgitation, type 2 diabetes with or without diabetic complications, congestive heart failure, suspected disease, angina pectoris, hypothyroidism, hyperphosphatemia, heart failure, renal anemia, sleep apnea syndrome, amplified valvular regurgitation, dehydration, hypertrophic cardiomyopathy, arrhythmia, peripheral neuropathy, cholinergic urticaria, plasma cell myeloma, chronic heart failure, and chronic renal failure.

42. 37. The method of claim 36, wherein the tests in the test history include a creatinine test.

43. 37. The method of claim 36, wherein the attributes include attributes selected from the group consisting of sex, age, age at diagnosis, and number of suspected diseases.

44. 37. The method of claim 36, wherein the machine learning model is selected from the group consisting of a decision tree, a random forest, a LightGBM, stacking, logistic regression, lasso regression, a support vector machine, a multilayer perceptron, a neural network, and combinations thereof.

45. A method for identifying hospitals where patients with Fabry disease may be treated, comprising the steps of: identifying patients who are predicted to have a possibility of having Fabry disease using the method described in any one of claims 33 to 44; identifying the area where the hospital where the patient is treated is located from the attribute information of the identified patients; and identifying hospitals located in the identified area.

46. 1. A computer program for predicting the likelihood that a patient has Fabry disease, the computer program comprising, when implemented on a computer: Inputting at least a portion of the patient's real-world data, including receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from a wearable device, into a predictive model; and obtaining an output from the predictive model indicating the likelihood that the patient has Fabry disease. A program containing instructions that cause a processor to execute the program.

47. A computer program product for causing a computer to carry out the method of any one of claims 33 to 44.

48. A computer-readable recording medium having the computer program according to claim 47 recorded thereon.

49. A trained machine learning model used to cause a computer to perform the method of any one of claims 35 to 44.

50. A method for generating a trained machine learning model for use in predicting the likelihood that a patient has Fabry disease, the method using teacher data for training, with at least a portion of the patient's real-world data, including prescription data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data from wearable devices, as explanatory variables, and the presence or absence of Fabry disease or suspicion of Fabry disease as the objective variable.

51. 51. The method for generating a trained machine learning model of claim 50, wherein at least a portion of the patient's real-world data includes the patient's diagnosis history, examination history, prescription history, medical procedure history, and / or attributes.

52. 1. A device for predicting the likelihood that a patient has Fabry disease, comprising: an input unit for inputting at least a portion of the patient's real-world data, including receipt data, DPC data, electronic medical record data, health checkup data, patient registry data, and / or data derived from a wearable device, into the prediction model; a prediction portion for predicting the likelihood that a patient has Fabry disease using a predictive model; and an output section for outputting a value indicating the likelihood that the predicted patient has Fabry disease; An apparatus having:

53. Apparatus comprising means for carrying out the method of any one of claims 33 to 44.

54. 48. An apparatus comprising a processor coupled to a memory device having stored thereon a computer program according to claim 47, the processor being capable of executing instructions of said program.

55. 50. An apparatus for predicting the likelihood that a patient has Fabry disease, comprising a prediction unit that makes predictions using the trained machine learning model described in claim 49.

56. 1. A method for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease, comprising: A step of inputting data including the doctor's gender, age, place of origin, alma mater, year of graduation, work facility history, medical department history, access behavior history of digital content related to Fabry disease, purchase history of medication related to Fabry disease, and / or paper writing history related to Fabry disease into a prediction model; and obtaining an output from the predictive model that indicates the likelihood that a physician will be able to diagnose and / or treat Fabry disease. A method comprising:

57. 57. The method of claim 56, wherein the predictive model makes a prediction based on the number or frequency of accesses to digital content related to Fabry disease.

58. The method of claim 56, wherein the value indicating the likelihood that a physician will be able to diagnose and / or treat Fabry disease is a predicted value by a prediction model or a stratified value of the predicted value.

59. The method of claim 58, wherein the predictive model is a machine learning model trained to output a value indicating the likelihood that the doctor will be able to diagnose and / or treat Fabry disease in response to input of one or more explanatory variables including the doctor's gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, history of access behavior to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers on Fabry disease.

60. The method of claim 59, wherein the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, access behavior history of digital content related to Fabry disease, purchase history of medications related to Fabry disease, and / or history of writing papers on Fabry disease of physicians known to be able to diagnose and / or treat Fabry disease.

61. The method of claim 59, wherein the predictive model is a machine learning model trained using one or more explanatory variables including the gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department, access behavior history of digital content related to Fabry disease, purchase history of medications related to Fabry disease, and / or history of writing papers on Fabry disease of doctors known to be unable to diagnose and / or treat Fabry disease.

62. 60. The method of claim 59, wherein the machine learning model is selected from the group consisting of a decision tree, a random forest, a LightGBM, stacking, logistic regression, lasso regression, a support vector machine, a multilayer perceptron, a neural network, and combinations thereof.

63. A method for identifying hospitals where doctors who can diagnose and / or treat Fabry disease may be employed, the method comprising the steps of identifying doctors who are predicted to have the potential to diagnose and / or treat Fabry disease using a method described in any one of claims 56 to 62, and identifying the hospital to which the doctors are employed from the attribute information of the identified doctors.

64. 1. A computer program for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease, which, when implemented on a computer, comprises: A step of inputting data including the doctor's gender, age, place of origin, alma mater, year of graduation, work facility history, medical department history, access behavior history of digital content related to Fabry disease, purchase history of medication related to Fabry disease, and / or paper writing history related to Fabry disease into a prediction model; and obtaining an output from the predictive model that indicates the likelihood that a physician will be able to diagnose and / or treat Fabry disease. A program containing instructions that cause a processor to execute the program.

65. The program of claim 64, wherein the predictive model makes a prediction based on the number or frequency of accesses to digital content related to Fabry disease.

66. A computer program product for causing a computer to carry out the method of any one of claims 56 to 62.

67. A computer-readable recording medium having the computer program of claim 66 recorded thereon.

68. A trained machine learning model used to cause a computer to perform the method of any one of claims 59 to 62.

69. A method for generating a trained machine learning model to be used to predict the likelihood that a doctor will be able to diagnose and / or treat Fabry disease, the method using training data to train a trained machine learning model in which the doctor's gender, age, place of origin, alma mater, year of graduation, history of working facility, history of medical department, history of access to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers on Fabry disease are used as explanatory variables, and whether or not the doctor will be able to diagnose and / or treat Fabry disease is used as the objective variable.

70. 1. A device for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease, comprising: an input section for inputting data into the prediction model, including the doctor's gender, age, place of origin, alma mater, year of graduation, history of facility of employment, history of medical department to which the doctor belongs, history of access to digital content related to Fabry disease, history of purchasing medications related to Fabry disease, and / or history of writing papers related to Fabry disease; a prediction component for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease using a predictive model; and an output unit for outputting a value indicative of the predicted likelihood that the physician will be able to diagnose and / or treat Fabry disease; An apparatus having:

71. 71. The apparatus of claim 70, wherein the predictive model makes a prediction based on the number or frequency of accesses to digital content related to Fabry disease.

72. 63. Apparatus comprising means for carrying out the method of any one of claims 56 to 62.

73. 65. An apparatus comprising a processor coupled to a memory device having stored thereon a computer program according to claim 64, the processor being capable of executing instructions of said program.

74. An apparatus for predicting the likelihood that a physician will be able to diagnose and / or treat Fabry disease, comprising a prediction unit that makes predictions using the trained machine learning model described in claim 68.

Images