Diagnosis support method, diagnosis support program, and diagnosis support system

Abstract

Provided is a diagnosis support system that acquires one or more items of patient information associated with a patient (steps S30, S40), and inputs the one or more items into an estimation model and acquires, from a plurality of diseases, a candidate for a disease that the patient may have (step S50). The diagnosis support system then acquires additional required information to be additionally acquired from the patient on the basis of the candidate for the disease (step S70). The diagnosis support system then presents the additional required information (step S90).

Description

Medical support method, medical support program, and medical support system

[0001] The present invention relates to medical assistance.

[0002] In particular, in small and medium-sized hospitals and clinics, there are cases where non-infectious disease specialists are required to treat infectious diseases. It is difficult for non-infectious disease specialists to become familiar with all infectious diseases, so there is a great need for clinical support from infectious disease specialists. Furthermore, with the increase in drug-resistant pathogens, the importance of appropriate use of antimicrobial drugs is becoming recognized, and from this perspective, there are also high expectations for clinical support from non-infectious disease specialists.

[0003] Compared to the number of medical institutions, there is an absolute shortage of infectious disease specialists who are well versed in infectious disease treatment, and there is a great demand for infectious disease treatment support tools. As an example of an infectious disease treatment support tool, a system using a database compiling information on infectious diseases has been proposed (see Non-Patent Document 1). When patient information, such as whether or not a patient has a cough or fever, is entered in this system, a list of infectious diseases that the patient may be suffering from is displayed, and information about the infectious diseases is also provided.

[0004] “Advancing the global effort against infectious diseases”, [online], [accessed May 10, 2022], Internet <URL: https: / / www.gideononline.com / >

[0005] According to the above system, even a doctor who is not a specialist in a given disease such as an infectious disease can know the possible diseases that a patient may be suffering from, making it easier to provide medical treatment.

[0006] However, the disease information output by this system is identified from limited patient information, and there is a need for a system that can output more reliable information.

[0007] The present invention has been devised in view of the above circumstances, and its purpose is to provide a technique for assisting non-specialists in diagnosing a disease in diagnosing that disease.

[0008] A medical support method according to an aspect of the present disclosure includes the steps of: acquiring one or more items of patient information related to the patient; inputting the one or more items into an estimation model to acquire, from a plurality of diseases, candidate diseases that the patient may be suffering from; obtaining additional request information that should be acquired from the patient based on the candidate diseases; and presenting the additional request information.

[0009] A medical assistance program according to an aspect of the present disclosure is executed by one or more processors of a computer, causing the computer to perform the above-described medical assistance method.

[0010] A medical support system according to an aspect of the present disclosure includes a processor, a communication interface, and a storage device that stores an estimation model configured to output, in response to input of one or more items of patient information, candidates for diseases from among a plurality of diseases that a person associated with the patient information may be suffering from, wherein the processor obtains input of values ​​for one or more items of patient information related to the patient, and obtains candidates for diseases from which the patient may be suffering by inputting the values ​​of the one or more items into the estimation model, and obtains additional request information that should be obtained from the patient based on the candidate diseases, and the communication interface presents the additional request information.

[0011] According to one aspect of the present disclosure, a non-specialist in a disease is provided with suggestions as to what item values ​​should be further considered as additional required information, thereby providing support to the non-specialist in a disease regarding the diagnosis and treatment of the disease.

[0012] 1 is a diagram for explaining the provision of information by the medical support system 100. FIG. 2 is a diagram illustrating an example of generated structured data. FIG. 3 is a diagram illustrating the hardware configuration of the medical support system 100. FIG. 4 is a diagram illustrating a mathematical formula representing an example of an estimation model. FIG. 5 is a flowchart of processing performed to present to the attending physician candidate infectious diseases that the patient may be suffering from. FIG. 6 is a diagram illustrating an example of an initial screen. FIG. 7 is a diagram illustrating an example of a screen displaying electronic medical record data input by the attending physician. FIG. 8 is a diagram illustrating an example of a screen after updating in accordance with the instruction of step S90. FIG. 9 is a diagram illustrating an example of a screen on which new physical findings have been input. FIG. 10 is a diagram illustrating an example of a new display screen.

[0013] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. In the drawings, the same or corresponding parts are designated by the same reference numerals, and description thereof will not be repeated.

[0014] [Overall Configuration] Fig. 1 is a diagram illustrating the provision of information by a medical assistance system 100. In one implementation example, a medical assistance system according to the present disclosure provides information regarding the diagnosis and / or treatment of a disease. In the following, an "infectious disease" will be described as an example of a disease. Note that the subject of the disease in the present disclosure is not limited to infectious diseases, and may also be a non-infectious disease.

[0015] 1 , the attending physician is a non-infectious disease specialist who examines a patient who is not sure whether or not the patient has an infectious disease, and operates the information terminal 500. The information terminal 500 is capable of communicating with the medical assistance system 100.

[0016] When electronic medical record data is input from the information terminal 500, the medical assistance system 100 analyzes the electronic medical record data to generate structured data. The structured data includes one or more items related to the patient and values ​​of each of the one or more items.

[0017] 2 is a diagram showing an example of the generated structured data. In the example shown in Fig. 2, the structured data includes at least five items. The structured data also includes values ​​for each of the five items shown in Fig. 2.

[0018] An example of an item is "fever," and an example of a value is "yes" for the item "fever." In one implementation, the value of the item "fever" in the structured data for a patient is "yes" if the patient's body temperature is 37.0°C or higher, and "no" if the body temperature is less than 37.0°C.

[0019] Returning to FIG. 1 , the medical support system 100 uses one or more values ​​in the structured data to identify infectious disease candidates that the patient may be suffering from. The medical support system 100 acquires medical reference information for each of the identified infectious disease candidates from the medical support information database. The medical support system 100 then transmits the infectious disease candidates and the medical reference information for each candidate to the information terminal 500.

[0020] The medical assistance system 100 identifies items whose values ​​are required to improve the accuracy of the identified candidates as "missing items." The medical assistance system 100 identifies the missing items based on the structured data, and then transmits the identified missing items to the information terminal 500.

[0021] The information terminal 500 displays the missing items. In response to this, the attending physician inputs the values ​​of the missing items into the information terminal 500. As a result, the values ​​of the missing items are added to the electronic medical record data.

[0022] The information terminal 500 transmits the electronic medical record data to which the values ​​of the missing items have been added to the medical assistance system 100. The medical assistance system 100 uses the electronic medical record data to which the values ​​of the missing items have been added to generate structured data, and again uses the generated structured data to identify candidates for infectious diseases that the patient may be infected with.

[0023] [Hardware Configuration] FIG. 3 is a diagram showing the hardware configuration of the medical assistance system 100. As shown in FIG.

[0024] The medical assistance system 100 is configured, for example, based on a personal computer. The medical assistance system 100 may also be configured as a server that can be accessed from one or more information terminals via a network such as the Internet.

[0025] The medical assistance system 100 includes a processor 101, a memory 200, and an input / output port 300. The input / output port 300 is connected to a mouse 400, a keyboard 120, and a display device 130.

[0026] The input / output port 300 may be a communication interface for data communication via a network. Patient data is input to the input / output port 300. The patient data is, for example, electronic medical record data. In one implementation example, structured data is generated from the electronic medical record data. The structured data may also be input to the input / output port 300. That is, the attending physician may extract the values ​​of one or more items that make up the structured data from the electronic medical record data, and input the values ​​of the one or more items to the input / output port 300 via the information terminal 500.

[0027] The memory 200 stores patient data 210, training data 220, an estimation model 230, a correspondence database 240, a medical support information database 250, and a support program 260.

[0028] The training data 220 is data relating to a plurality of individuals and is used for machine learning of the estimation model 230. In the training data 220, structured data for each of the individuals, as described with reference to FIG. 2 , is tagged with information about the infectious disease contracted by each individual. The information about the contraction includes whether or not the individual has an infectious disease and the type of infectious disease contracted.

[0029] The estimation model 230 is a program for performing calculations according to the model. An example of the algorithm of the estimation model 230 is logistic regression, but the algorithm followed by the estimation model 230 is not limited to this. In one implementation example, the estimation model 230 is prepared for each of two or more types of infectious diseases, and the estimation model for each infectious disease derives whether or not a patient is infected with the infectious disease. Note that the estimation model for each infectious disease may also derive a probability (likelihood).

[0030] The correspondence database 240 is an example of correspondence information, and associates each of two or more types of infectious diseases with an item (one or more items included in the structured data) used for estimation by the estimation model 230. For example, the correspondence database 240 associates the infectious disease "pyelonephritis" with three types of items (fever, frequent urination, and flank pain).

[0031] The medical support information database 250 associates each of two or more types of infectious diseases with medical support information (information for supporting doctors in medical treatment) for each infectious disease. The medical support information is an example of additional information for an infectious disease.

[0032] The assistance program 260 includes, as its functions, a data analysis unit 261, a model generation unit 262, a first information generation unit 263, a data collection unit 264, a second information generation unit 265, and an output unit 266. In one implementation example, the functions of the data analysis unit 261, the model generation unit 262, the first information generation unit 263, the data collection unit 264, the second information generation unit 265, and the output unit 266 are realized by the processor 101 executing a given program.

[0033] The data analysis unit 261 creates structured data such as that described with reference to FIG. 2 from the patient data 210 .

[0034] In one implementation example, the data analysis unit 261 applies an analysis process such as morphological analysis to the text of the electronic medical record data stored as the patient data 210, thereby extracting portions of the text that constitute a pre-registered pattern. The "pre-registered pattern" is a pattern that corresponds to a value of the structured data. The data analysis unit 261 then recognizes the meaning of the portion that constitutes the extracted pattern in the electronic medical record data and identifies a value that corresponds to the recognized meaning. The data analysis unit 261 then generates structured data by combining one or more of the identified values.

[0035] An example of a registered pattern is "body temperature is xx°C," where "xx" represents a numerical value. If the electronic medical record data includes the text "body temperature is 37.5°C," the data analysis unit 261 extracts the text "body temperature is 37.5°C" as a component of the pre-registered pattern. The data analysis unit 261 then recognizes that the meaning of the text is that the patient's body temperature is 37.5°C. Meanwhile, the data analysis unit 261 is configured to identify items and their values ​​in the structured data based on the recognized meaning. For example, if the body temperature is 37.0°C or higher, the item "fever" and its value "present" are identified, and if the body temperature is less than 37.0°C, the item "fever" and its value "absent" are identified. Thus, if the electronic medical record data includes the text "body temperature is 37.5°C," the data analysis unit 261 generates structured data having the value "present" for the item "fever."

[0036] The structured data items include at least one of medical history information, physical examination findings, and test results. The medical history information, physical examination findings, and test results may be included in the electronic medical record data.

[0037] The medical history information includes information obtained from the patient by the attending physician during the patient interview. The medical history information may also include information obtained from the patient by another physician during the patient interview in the past. An example of an item included in the medical history information is the patient's chief complaint.

[0038] Physical examination findings include information collected from the patient by the attending physician during examination (inspection, auscultation, palpation, percussion, tendon reflexes, light reflexes, etc.).

[0039] The test results include the results of tests performed on the patient (blood tests, CT (Computed Tomography) scans, temperature measurements, weight measurements, blood pressure measurements, heart rate measurements, etc.).

[0040] The structured data items may also include basic information about the patient (age, sex, level of care required, smoking history, etc.), allergy information, and / or items related to medical history.

[0041] The model generation unit 262 uses the training data 220 to perform machine learning processing on the estimation model 230. In this specification, the estimation model 230 that has been subjected to machine learning processing is also referred to as a "trained model."

[0042] The first information generation unit 263 derives a determination result regarding an infectious disease that the patient may be suffering from by applying the structured data generated from the patient data to the trained model. In one implementation example, the first information generation unit 263 uses the trained model for each type of infectious disease to derive the probability (likelihood) that the patient is suffering from each infectious disease, and identifies an infectious disease for which the probability is equal to or greater than a given value as a candidate infectious disease.

[0043] The data collection unit 264 collects medical support information for each of the candidates identified by the first information generation unit 263 from the medical support information database 250 .

[0044] The second information generation unit 265 refers to the correspondence database 240 to identify, for each of the candidates identified by the first information generation unit 263, items that are missing values ​​in the structured data generated from the patient data 210 as missing items.

[0045] The output unit 266 generates image information for displaying the information specified in the assistance program 260 and instructs the information terminal 500 to display the image information.

[0046] [Identification of infectious disease candidates] A description will be given of an implementation example of identifying infectious disease candidates by the first information generation unit 263. As described above with reference to Fig. 3 , in one implementation example, an estimation model 230 is prepared for each type of infectious disease, and the first information generation unit 263 applies structured data to each estimation model 230 to derive the probability that the patient has each infectious disease.

[0047] FIG. 4 is a diagram showing a mathematical formula representing an example of an estimation model. Equation (1) shown in FIG. 4 follows logistic regression analysis. In equation (1), the subscript represents the type of item included in the structured data. In equation (1), i types of items are specified as "1" to "i." The value of "i" may differ depending on the type of infectious disease. In other words, the type and number of items used to estimate the probability of contracting each infectious disease may differ for each infectious disease. The correspondence database 240 associates each infectious disease with the i types of items included in equation (1) prepared for that infectious disease.

[0048] In formula (1), x is an explanatory variable, and corresponds to the value of each item of the structured data. For example, for the item "fever," if the value is "present," x is 1, and if the value is "absent," x is 0.

[0049] In equation (1), b is a coefficient. ０ ~b ｉ The values ​​of each are set in machine learning for each infectious disease.

[0050] In equation (1), y is a response variable that represents the calculation result for each infectious disease. The value of y represents the likelihood of contracting each infectious disease.

[0051] The first information generation unit 263 derives the likelihood that the patient is infected with each infectious disease by applying the value of each item included in the structured data to the estimation model for each infectious disease.

[0052] Then, the first information generation unit 263 identifies infectious diseases whose "likelihood" value is equal to or greater than a given threshold as candidates for the infectious disease that the patient is suffering from.

[0053] The manner in which infectious disease candidates are identified as described with reference to Figure 4 is merely an example. Infectious disease candidates may be identified in any other manner, such as multivariate analysis using one or more types included in the structured data as variables.

[0054] Furthermore, when structured data is applied to an estimation model, for some infectious diseases, the structured data may not include all values ​​of the items defined in the estimation model, i.e., some of the items defined in the estimation model may be missing from the structured data.

[0055] For example, assume that the estimation model for the infectious disease "pyelonephritis" defines three items (fever, frequent urination, and flank pain). If the attending physician includes information on fever and frequent urination but not on flank pain in the electronic medical record data, the structured data will include values ​​for the items "fever" and "frequent urination" but not the item "flank pain." In such a case, the first information generating unit 263 performs missing value processing when deriving the likelihood using the estimation model. One example of missing value processing is randomly filling in the value of a missing item with the values ​​of other items used in the formula of the same estimation model. Another example is filling in the value of a missing item with the average value of the values ​​of other items used in the formula of the same estimation model. For example, if the value of the item "flank pain" is missing, this value is filled in with the average value of the values ​​of the items "fever" and "frequent urination."

[0056] [Identification of Missing Items] The second information generation unit 265 determines whether or not there are any missing items as described above for each infectious disease identified as a candidate by the first information generation unit 263. More specifically, for each infectious disease identified as a candidate by the first information generation unit 263, if there is any item associated with the infectious disease in the correspondence database 240 that is not included in the structured data, the second information generation unit 265 determines that there are any missing items. Furthermore, if all of the items associated with the infectious disease in the correspondence database 240 are included in the structured data, the second information generation unit 265 determines that there are no missing items.

[0057] If there is a missing item, the second information generating unit 265 identifies the missing item as a missing item.

[0058] [Important Factors] The importance score of each item calculated when the estimation model was generated for each infectious disease may be stored in the memory 200. The second information generation unit 265 may identify, for each infectious disease, items included in the structured data that have an importance score equal to or greater than a given value as important factors.

[0059] "Above a given value" may mean that the importance score has a "positive" value above a certain value.

[0060] "Above a given value" may mean that the absolute value of the importance score is above a given value. When the importance score of an item is a positive value, the magnitude of the absolute value of the score represents the likelihood of an infectious disease. When the importance score of an item is a negative value, the magnitude of the absolute value of the score represents the unlikelihood of an infectious disease. When "above a given value" means that the absolute value of the importance score is above a given value, the item identified as an important factor characterizes the likelihood of an infectious disease or the unlikelihood of an infectious disease.

[0061] [Processing Flow] Figure 5 is a flowchart of processing performed to present to the attending physician candidate infectious diseases that the patient may be suffering from. In one implementation example, the processing of Figure 5 is realized by the processor 101 of the medical assistance system 100 executing a given program (medical assistance program). The processing of Figure 5 will be described below. In one implementation example, the processing of Figure 5 is started in response to the medical assistance system 100 receiving a request from the information terminal 500 to start providing information. In one implementation example, the information terminal 500 requests the medical assistance system 100 to start providing information in response to the launch of a medical assistance application program and an operation being performed in the application program to request the start of information provision.

[0062] In step S10, the medical assistance system 100 instructs the information terminal 500 to display an initial screen. As a result, the information terminal 500 displays the initial screen on the display of the information terminal 500.

[0063] 6 shows an example of an initial screen 600. In FIG. 6, a screen 600 includes a first communication frame 610, fields 621 to 627 and 630, and a second communication frame 640.

[0064] The first communication frame 610 displays information for the attending physician to communicate with the medical support system 100. The first communication frame 610 includes a field 611 and a button 619. In the example of Fig. 6, a message to the attending physician (Please enter the medical record) is displayed in the field 611. The button 619 is operated to instruct the medical support system 100 to provide information regarding the data entered on the screen.

[0065] Each of columns 621 to 627 contains the patient's electronic medical record data. More specifically, column 621 contains basic patient information (e.g., age, gender, level of care required, smoking history, etc.). Column 622 contains the patient's medical history. Column 623 contains the patient's allergy information. Column 624 contains vital signs from the results of tests performed on the patient. Column 625 contains test results other than vital signs (e.g., blood test results, CT scan results, etc.). Column 626 contains physical findings. Column 627 contains medical history information (e.g., chief complaint). Note that the input format of the electronic medical record data described above is merely an example. The medical support system 100 may receive electronic medical record data written in text format and generate the screen data shown in FIG. 6 by extracting the data to be displayed in columns 621 to 627 from the electronic medical record data.

[0066] In the field 630, information on possible infectious diseases that the patient may be suffering from is displayed, which is provided to the attending physician by the medical support system 100.

[0067] The second communication frame 640 is used by the attending physician to communicate with the infectious disease specialist. The attending physician, viewing the screen 600, inputs electronic medical record data into fields 621 to 627. Figure 7 is a diagram showing an example of a screen displaying the electronic medical record data input by the attending physician. In Figure 7, the screen 601 displays the electronic medical record data input by the attending physician in each of fields 621 to 627.

[0068] Thereafter, the attending physician operates the button 619. As a result, the input electronic medical record data is transmitted from the information terminal 500 to the medical assistance system 100.

[0069] 5, in step S20, the medical assistance system 100 determines whether or not an instruction to provide information has been received from the information terminal 500. The medical assistance system 100 repeats the control of step S20 until it determines that an instruction to provide information has been received (NO in step S20). When the medical assistance system 100 determines that an instruction to provide information has been received (YES in step S20), the control proceeds to step S30.

[0070] In step S30, the medical assistance system 100 reads the electronic medical record data transmitted from the information terminal 500.

[0071] In step S40, the medical assistance system 100 generates structured data as described with reference to FIG. 2 using the electronic medical record data read in step S30.

[0072] In step S50, the medical support system 100 uses the structured data generated in step S40 to identify one or more candidate infectious diseases that the patient may be suffering from.

[0073] In step S60, the medical support system 100 refers to the medical support information database 250 to acquire medical support information for each of the one or more infectious disease candidates identified in step S50.

[0074] In step S70, the medical assistance system 100 identifies missing items for each of the one or more infectious disease candidates identified in step S50.

[0075] In step S80, the medical support system 100 identifies important factors for each of the one or more infectious disease candidates identified in step S50.

[0076] In step S90, the medical assistance system 100 uses the results of steps S50 to S80 to generate a display screen that displays information such as suspected infectious diseases, and instructs the information terminal 500 to display the generated display screen. This updates the display screen on the information terminal 500. The medical assistance system 100 then returns control to step S20.

[0077] 8 is a diagram showing an example of a screen after updating in response to the instruction in step S90. In Fig. 8, a screen 602 is an example of a screen after updating the screen 601 in Fig. 7.

[0078] In comparison with the screen 601 of Fig. 7, the screen 602 of Fig. 8 shows one or more candidates that the patient may be suffering from in a field 630. Each of the one or more tabs displayed in the field 630 corresponds to one or more candidate infectious diseases identified in step S50.

[0079] Column 630 of FIG. 8 shows two tabs representing two types of infections: pyelonephritis and urinary tract infections.

[0080] More specifically, the column 630 displays two tabs so that the tab displaying the contents can be switched. Fig. 8 shows a state in which the contents of the "pyelonephritis" tab are displayed, out of the "pyelonephritis" tab and the "urinary tract infection" tab. The attending physician can display the contents of the "urinary tract infection" tab in the column 630 by operating the information terminal 500.

[0081] The contents of each tab in the column 630 are medical support information for the corresponding infectious disease. The medical support information for each infectious disease candidate acquired in step S60 is displayed in each tab in the column 630.

[0082] Compared to FIG. 7 , the first communication frame 610 in FIG. 8 has an additional column 612. When a missing item is identified in step S70, the medical assistance system 100 adds a column to the first communication frame 610 and displays a message in the added column prompting the user to enter a value for the missing item. A column may be added for each missing item. For example, when two types of candidates are identified and one missing item is identified for each of the two types of candidates, two columns may be added. A message prompting the user to enter a value for the missing item identified for the first candidate may be displayed in one of the two added columns, and a message prompting the user to enter a value for the missing item identified for the second candidate may be displayed in the other of the two added columns.

[0083] In the example of Fig. 8, the message "Please enter findings related to flank pain" is displayed in the field 612. This message corresponds to the identification of "flank pain" as a missing item. More specifically, the medical support system 100 displays this message to prompt the attending physician to further input values ​​for the identified missing items (as physical findings).

[0084] In the example of Figure 8, compared to Figure 7, some of the text displayed in columns 621 to 627 are highlighted by being surrounded by a box. The highlighted text corresponds to the important factors identified in step S80. In one implementation example, the important factors are highlighted as reasons for identifying an infectious disease as a candidate.

[0085] More specifically, the medical support system 100 highlights the portions of the text (electronic medical record data) displayed in columns 621 to 627 that correspond to the important factors identified in step S80. The portions that correspond to the important factors are the portions of the electronic medical record data that were used to generate the values ​​of the items identified as important factors in generating the structured data.

[0086] For example, assume that the infectious disease "pyelonephritis" is identified as a candidate, the correspondence database 240 defines three types of items (fever, frequent urination, and flank pain) as items used to estimate the infectious disease "pyelonephritis," and of these, "fever" and "frequent urination" are identified as important factors. In this case, the structured data further includes the value "present" of the item "fever" generated from the text "body temperature 37.2°C" in the electronic medical record data, and also includes the value "present" of the item "frequent urination" generated from the text "increased frequency of urination" in the electronic medical record data.

[0087] In the above case, the part of the electronic medical record data that corresponds to the important factor "fever" is "body temperature 37.2°C," and the part that corresponds to the important factor "frequent urination" is "increased frequency of urination." Therefore, in the above case, in columns 621 to 627, "body temperature 37.2°C" and "increased frequency of urination" are highlighted and displayed in the electronic medical record data.

[0088] 5, when the attending physician inputs the patient's electronic medical record data, candidate infectious diseases that the patient may have are presented to the attending physician, and items that are considered to be missing in order to identify the candidate as the infectious disease that the patient has are presented to the attending physician as a message prompting the physician to enter values ​​for the missing items. This provides the attending physician with suggestions as to what information on the presented candidate is needed to use the missing items in a final diagnosis.

[0089] The attending physician may see the message and additionally input values ​​for the missing items. The medical support system 100 may again identify candidates for infectious diseases that the patient may be infected with, using the initially input values ​​and the additionally input values ​​for the missing items.

[0090] More specifically, when the medical assistance system 100 instructs the information terminal 500 to update the screen in step S90, the information terminal 500 displays a message prompting the user to enter values ​​for the missing items, as shown in field 612 in Fig. 8. The attending physician inputs new physical findings into the screen in accordance with the message displayed in field 612.

[0091] 9 is a diagram showing an example of a screen on which new physical findings have been entered. Compared to screen 602 in FIG. 8, screen 603 in FIG. 9 further includes, in field 626, text corresponding to the new physical findings: "Abdomen_Distention_Crying 'Ouch!' when lightly pressed on both flanks." The attending physician then operates button 619.

[0092] After issuing an instruction to update the screen in step S90, the medical assistance system 100 returns control to step S20. In response to the attending physician operating button 619, the medical assistance system 100 determines that an instruction to provide information has been issued (YES in step S20), and proceeds to step S30.

[0093] In step S30, the medical support system 100 reads the electronic medical record data including the new physical findings. Then, in step S40, the medical support system 100 generates new structured data using the electronic medical record data read in step S30. Then, in step S50, the medical support system 100 identifies an infectious disease candidate using the new structured data. In step S60, the medical support system 100 acquires medical support information for the candidate identified in step S50. In step S70, the medical support system 100 identifies missing items for the candidate identified in step S50. In step S80, the medical support system 100 identifies important factors for the candidate identified in step S50. In step S90, the medical support system 100 generates a new display screen using the results of steps S50 to S80 performed after reading the electronic medical record data including the new physical findings, and instructs the information terminal 500 to display the new display screen. This updates the display screen on the information terminal 500.

[0094] Fig. 10 is a diagram showing an example of a new display screen. In screen 604 of Fig. 10, compared to screen 603 of Fig. 9, field 630 includes only "pyelonephritis" as a candidate infectious disease. Also, compared to screen 603 of Fig. 9, screen 604 of Fig. 10 adds the text "It hurts just by lightly pressing the left and right flanks" in field 626 as emphasized text.

[0095] In the example described with reference to Figures 9 and 10, the value "Yes" of the item "Flank Pain" was added to the structured data based on the text "It hurts when I just lightly press on the left and right flanks," and the likelihood of the urinary tract infection, out of the two candidates (pyelonephritis and urinary tract infection) shown in Figure 8, decreased. That is, in this example, the number of infectious disease candidates decreased due to the input of the missing item value. This allows for more accurate information to be provided to the patient's doctor as possible infectious disease candidates.

[0096] In this embodiment, medical support information is also provided for each candidate, so that even if the attending physician is not an infectious disease specialist, the attending physician can be provided with information that is useful for treating the patient.

[0097] In this embodiment, for each candidate, the portion of the electronic medical record data corresponding to the important factors is highlighted, which provides the attending physician with useful information when making a final diagnosis using the presented candidates.

[0098] In this embodiment, the medical support information database 250 is stored in the memory 200 of the medical support system 100, but it may be stored in a storage device other than the medical support system 100. The medical support system 100 may access the storage device to acquire medical support information for each of one or more infectious disease candidates.

[0099] (Variation) In this embodiment, a medical care support system has been described that identifies missing items that, if input into an estimation model, can improve the accuracy of infectious disease candidates, and then presents the missing items to the attending physician. Note that in this embodiment, the missing items are an example of information that should be additionally obtained from the patient (additional request information). When obtaining the missing items, a correspondence database (an example of correspondence information) is referenced.

[0100] In this embodiment, in step S90, information based on a request from a specialist may be presented instead of or in addition to the missing items. The information based on a request from a specialist is, for example, information of a type required by the specialist for the type of infectious disease identified as a candidate, and is information of a type that was not included in the structured data generated in step S40 (the electronic medical record data read in step S30). In this context, the information based on a request from a specialist is another example of additional requested information.

[0101] The types of information required by the specialist may be stored in advance in a storage device such as memory 200 for each type of infectious disease, for example, as a request correspondence table. In this context, the request correspondence table constitutes an example of information describing the relationship between each of multiple infectious diseases and additional request information. In one implementation example, once a candidate infectious disease is estimated using the estimation model, the medical assistance system 100 reads the specialist's request corresponding to the candidate infectious disease from the storage device using the request correspondence table. Then, among one or more types of information included in the read request, information of a type not included in the structured data generated in step S40 is identified as information based on the specialist's request. This allows the attending physician to obtain the type of information required by the specialist for the candidate infectious disease in advance (before directly contacting the specialist). This allows the attending physician to respond to the patient more quickly by reducing the number and / or time of interactions with the specialist.

[0102] It should be noted that the identification of information based on the specialist's request may be performed in addition to or instead of identifying the missing items in step S70.

[0103] In addition to or instead of referencing the request correspondence table, the identification of information based on the specialist's request may be performed using the response obtained by the attending physician from the specialist. More specifically, the attending physician may present the electronic medical record data and suspected infectious diseases to the specialist. In response, the specialist may request information necessary to identify infectious diseases that the patient may be suffering from in his or her response to the attending physician. The medical support system 100 may identify the "necessary information" included in the response from the specialist as the "information based on the specialist's request." The medical support system 100 may identify the "information based on the specialist's request" by performing natural language processing on the response from the specialist.

[0104] In one implementation example, the medical support system 100 may accept an operation to send information to a specialist on a display screen displaying information such as a suspected infectious disease in step S90. In response to the operation, the medical support system 100 may send the electronic medical record data and the suspected infectious disease to an address associated with the specialist. Thereafter, when the medical support system 100 receives a response from the specialist, it may update the display screen to add the response. This allows the response from the specialist to be provided to the attending physician. The information sent to the specialist may be a display screen generated in step S90. This allows the patient information, the suspected disease, and the attending physician's diagnosis to be sent to the specialist.

[0105] The medical support system may present both the information necessary for inferring infectious disease candidates using the estimation model and the specialist's requests to the attending physician. The attending physician may obtain the patient's findings while referring to the specialist's requests, and then additionally input the findings into the medical support system. The medical support system may use the additionally input information (findings) as described above for re-estimation using the estimation model. The content of the "information necessary for inferring infectious disease candidates using the estimation model" and the content of the "additionally input information" as described above may overlap. The medical support system may present these two types of information to the attending physician in a manner that is distinguishable from each other (for example, visually), or may present them in the same manner.

[0106] (Reference Data) "Reference data" may be used as information referenced in identifying missing items in step S70. The reference data stores information used for diagnosis or treatment for each disease. In step S70, the medical assistance system 100 identifies missing items for each of the one or more disease candidates identified in step S50 by referring to the reference data. In one implementation example, the medical assistance system 100 identifies, as missing items, types of information stored in the reference data for the disease identified in step S50 that are not included in the information already acquired.

[0107] In the "information used for diagnosis or treatment," information used for diagnosis and information used for treatment may be distinguished from each other. For example, for the disease "pyelonephritis," types of information used for diagnosis include "fever," "frequent urination," "flank pain," and "age," and types of information used for treatment include "medical history" and "age."

[0108] The instruction to provide information from the information terminal 500, which is the subject of the determination in step S20, may include the purpose of providing the information (diagnosis or treatment). In identifying the missing items in step S70, the medical assistance system 100 may use either the information used for diagnosis or the information used for treatment among the "information used for diagnosis or treatment" according to the purpose included in the instruction. More specifically, if the purpose is "diagnosis," the medical assistance system 100 may use the information used for diagnosis, and if the purpose is "treatment," the medical assistance system 100 may use the information used for treatment.

[0109] In the "information used for diagnosis or treatment," one or more types of information may be classified as "essential information" and "optional information." In the above-mentioned information used for diagnosing the disease "pyelonephritis" ("fever," "frequent urination," "flank pain," and "age"), "fever," "frequent urination," and "flank pain" may be classified as "essential information," while "age" may be classified as "optional information."

[0110] The missing items identified using the reference data may be output to the display screen described above with respect to step S90.

[0111] Furthermore, the display screen may include an indication indicating whether the information already acquired about the patient by the medical assistance system 100 includes all of the types of information identified as the above-mentioned "required information." One example of this indication is an "OK" mark that is displayed when the information already acquired about the patient by the medical assistance system 100 includes all of the above-mentioned "required information." Another example is an "NG" mark that is displayed when the information already acquired about the patient by the medical assistance system 100 does not include at least some of the above-mentioned "required information."

[0112] Another example of the display may show the percentage of the types of "essential information" that are included in the information that the medical support system 100 has already acquired about the patient, relative to all types of information. In one example, the display showing this percentage is a numerical value (100%, 50%, etc.). In another example, if there are three types of essential information ("fever," "frequent urination," and "flank pain"), the display showing the percentage will be "Good" if all three types are included, "OK" if one or two types are included, and "NG" if none are included.

[0113] (Display of guidelines) The medical assistance system 100 may add information extracted from the guidelines to the display screen generated in step S90. In one implementation example, the medical assistance system 100 adds diagnostic or treatment guidelines corresponding to the disease identified in step S50 to the display screen. An example of a guideline that can be added is the JAID / JSC Infectious Disease Treatment Guidelines (https: / / www.kansensho.or.jp / uploads / files / guidelines / guideline_JAID-JSC_2015_urinary-tract.pdf).

[0114] The medical support system 100 may extract a portion of the guideline that corresponds to the identified disease and add only the extracted portion to the display screen. The medical support system 100 may also add only either the guideline article itself or access information (e.g., a URL) to the display screen. Similarly, the medical support system 100 may also add information extracted from the guideline for diseases based on the attending physician's findings to the display screen.

[0115] The medical assistance system 100 may select the type of guideline to add according to the purpose included in the information provided from the information terminal 500 that is the subject of the determination in step S20. More specifically, the medical assistance system 100 may add information about diagnostic guidelines to the display screen if the purpose is "diagnosis," and may add information about treatment guidelines to the display screen if the purpose is "treatment."

[0116] Aspects It will be understood by those skilled in the art that the exemplary embodiments described above are examples of the following aspects.

[0117] (Section 1) A medical care support method according to one aspect may include the steps of acquiring one or more items of patient information related to a patient, inputting the one or more items into an estimation model to acquire, from a plurality of diseases, candidate diseases that the patient may be suffering from, acquiring additional request information that should be acquired from the patient based on the candidate diseases, and presenting the additional request information.

[0118] According to the medical care support method of paragraph 1, a non-specialist in a disease is provided with suggestions as to what item values ​​should be further considered as additional required information, thereby providing support for the non-specialist in a disease regarding the medical care of the disease.

[0119] (Clause 2) In the medical support method of clause 1, the step of obtaining the additional request information may include referring to reference data in which a disease is associated with information used for diagnosis or treatment, and identifying the additional request information based on the information used for diagnosis or treatment that is associated with the candidate disease.

[0120] According to the medical care support method of the second paragraph, information for examining the appropriateness of possible diseases and treatment methods can be provided to the attending physician.

[0121] (Clause 3) In the medical support method of paragraph 2, the reference data may include required information and optional information as information to be used for the diagnosis or treatment, and the method may further include a step of presenting a determination result as to whether the one or more items include all of the required information.

[0122] According to the medical care support method of the third aspect, the attending physician can be provided with information for considering whether or not it is actually necessary to input the additional request information.

[0123] (Clause 4) In the medical support method of clause 2 or clause 3, the information used for the diagnosis or treatment includes diagnostic information and therapeutic information, and the method further includes a step of receiving a diagnosis or treatment for the purpose of obtaining a candidate disease, and specifying the additional request information may include specifying the additional request information based on the diagnostic information if the purpose is diagnosis, and specifying the additional request information based on the therapeutic information if the purpose is treatment.

[0124] According to the medical care support method of the fourth aspect, appropriate information can be provided to the attending physician depending on the situation (treatment or diagnosis) the attending physician faces.

[0125] (Item 5) The medical assistance method according to any one of items 1 to 5 may further include a step of displaying a diagnostic guideline or a treatment guideline.

[0126] According to the medical care support method of the fifth aspect, more information regarding the disease can be provided to the attending physician.

[0127] (Clause 6) The medical support method of any one of clauses 1 to 5 may further include a step of accepting the diagnosis result of the attending physician together with the patient information, and a step of transmitting the patient information, the suspected disease, and the diagnosis result of the attending physician to a specialist.

[0128] According to the medical care support method of Section 6, more information is provided to the specialist, which enables the specialist to give a more appropriate answer.

[0129] (Clause 7) The diagnostic support method according to any one of clauses 1 to 5 may further include a step of highlighting the basis on which the estimation model has determined the disease candidate.

[0130] According to the medical care support method of paragraph 7, it is possible to provide a basis for judging inference, which is a black box. (paragraph 8) In the medical care support method of any one of paragraphs 1 to 7, the additional request information may include information that can improve the accuracy of the disease candidate if input to the estimation model, and the step of obtaining the additional request information may include obtaining the additional request information by referencing correspondence information that associates each of the plurality of diseases with an item of the patient information.

[0131] According to the medical care support method of the eighth aspect, the step of acquiring additional request information is more specifically provided.

[0132] (Clause 9) In the medical support method of any one of clauses 1 to 8, the additional request information may include missing items that are not included in the one or more items of patient information associated with the disease candidate, and the step of obtaining the additional request information may include identifying the missing items.

[0133] According to the medical care support method of the ninth aspect, more specific information is provided as additional request information.

[0134] (Clause 10) In the medical care support method of any one of clauses 1 to 9, the step of obtaining the additional request information may include obtaining the additional request information corresponding to the obtained candidate disease from a storage device that stores information describing the relationship between each of the plurality of diseases and the additional request information, or obtaining the additional request information by presenting the obtained candidate disease to a specialist and utilizing the response obtained.

[0135] According to the medical care support method of the tenth aspect, more specific information is provided as additional request information.

[0136] (Clause 11) In the medical support method of any one of clauses 1 to 10, the step of obtaining values ​​of the one or more items may include accepting input of electronic medical record data of the patient, and generating values ​​of the one or more items using the electronic medical record data.

[0137] According to the medical care support method of paragraph 11, in order to receive information, a non-specialist doctor only needs to input the patient's medical record data instead of inputting values ​​for one or more items. This makes the work of the non-specialist doctor easier.

[0138] (Clause 12) The medical support method of any one of clauses 1 to 11 further comprises a step of presenting the candidate disease, and the step of presenting the candidate disease may include presenting an item among the one or more items that has an importance equal to or greater than a given value in identifying the candidate disease, or additional information about the candidate disease.

[0139] According to the medical care support method of Section 12, information is provided that helps understand the reason why the presented candidates were presented.

[0140] (13) In the medical care support method according to any one of paragraphs 1 to 12, the patient information may include at least one of information on medical history, physical findings, and test results.

[0141] According to the medical care support method of paragraph 10, candidate infectious diseases that the patient may be suffering from are provided by inputting information available to the patient's doctor.

[0142] (Clause 14) A medical assistance program according to one aspect may be executed by one or more processors of a computer to cause the computer to implement the medical assistance method of any one of clauses 1 to 13.

[0143] According to the medical care support program of paragraph 14, non-specialists in a disease are provided with suggestions as to what item values ​​they should further consider as additional required information, thereby providing support for the non-specialists in the medical care of the disease.

[0144] (Clause 15) A medical assistance system according to one embodiment includes a processor, a communication interface, and a storage device that stores an estimation model configured to output, in response to input of one or more items of patient information, candidates for diseases that a person associated with the patient information may be suffering from among a plurality of diseases, wherein the processor obtains input of values ​​for one or more items of patient information associated with the patient, and inputs the values ​​of the one or more items into the estimation model to obtain candidates for diseases that the patient may be suffering from, and obtains additional request information that should be obtained from the patient based on the candidate diseases, and the communication interface may present the additional request information.

[0145] According to the medical care support system of paragraph 15, non-specialists in a disease are provided with suggestions as to what item values ​​they should further consider as additional required information, thereby providing support for the non-specialists in the diagnosis of the disease.

[0146] The embodiments disclosed herein should be considered to be illustrative in all respects and not restrictive. The scope of the present disclosure is defined by the claims, not by the description of the above-described embodiments, and is intended to include all modifications within the meaning and scope of the claims. Furthermore, it is intended that each technique in the embodiments can be implemented alone or, if necessary, in combination with other techniques in the embodiments to the extent possible.

[0147] 100 Medical support system, 101 Processor, 500 Information terminal, 600, 601, 602, 603, 604 Screen, 610 First communication frame, 611, 612, 621, 622, 623, 624, 625, 626, 627, 630 Column, 619 Button, 640 Second communication frame.

Claims

1. obtaining one or more items of patient information associated with the patient; inputting the one or more items into a prediction model to obtain candidates for a disease that the patient may be suffering from from among a plurality of diseases; obtaining additional required information to be obtained from the patient based on the disease candidate; and presenting the additional request information.

2. The step of obtaining additional request information includes: Referencing reference data in which a disease is associated with information to be used for diagnosis or treatment; The medical support method according to claim 1 , further comprising: identifying the additional request information based on information associated with the candidate disease and used for the diagnosis or treatment.

3. The reference data includes required information and optional information as information used for the diagnosis or treatment, The medical support method according to claim 2 , further comprising the step of presenting a result of a determination as to whether or not the one or more items include all of the required information.

4. The information used for diagnosis or treatment includes diagnostic information and therapeutic information, and administering the disease candidate to a patient for diagnosis or treatment, Identifying the additional request information includes: If the purpose is diagnosis, identifying the additional request information based on the diagnosis information; The medical support method according to claim 2 , further comprising: if the purpose is treatment, specifying the additional request information based on the treatment information.

5. The medical support method according to claim 1 , further comprising the step of displaying a diagnostic guideline or a treatment guideline for the candidate disease.

6. receiving a diagnosis of the attending physician together with the patient information; The medical support method according to claim 1 , further comprising the step of transmitting the patient information, the suspected disease, and the diagnosis of the attending physician to a specialist.

7. The medical support method according to claim 1 , further comprising the step of highlighting, in the patient information, a basis for the determination of the disease candidate by the estimation model.

8. the additional request information includes information that, when input to the estimation model, can improve the accuracy of the disease candidate; The medical support method according to claim 1 , wherein the step of obtaining the additional request information includes obtaining the additional request information by referring to correspondence information that associates each of the plurality of diseases with an item of the patient information.

9. The additional request information includes a missing item that is not included in the one or more items of patient information associated with the disease candidate, The medical support method according to claim 1 , wherein the step of obtaining the additional required information includes identifying the missing items.

10. The step of obtaining additional request information includes: Obtaining additional request information corresponding to the obtained disease candidate from a storage device that stores information describing a relationship between each of the plurality of diseases and additional request information; or 2. The method for medical support according to claim 1, further comprising: presenting the obtained candidate diseases to a specialist and obtaining additional required information using the obtained response.

11. The step of obtaining values ​​of one or more items includes: accepting input of electronic medical record data of the patient; The medical support method according to claim 1 , further comprising: generating values ​​for the one or more items using the electronic medical record data.

12. The method further comprises the step of presenting the candidate disease; The medical support method according to claim 1 , wherein the step of presenting the disease candidates includes presenting, among the one or more items, items that have an importance level equal to or greater than a given value in identifying the disease candidates or additional information on the disease candidates.

13. The medical support method according to claim 1 , wherein the patient information includes at least one of information on a medical interview, a physical examination, and a test result.

14. A medical assistance program that, when executed by one or more processors of a computer, causes the computer to carry out the medical assistance method according to claim 1.

15. A processor; A communication interface; a storage device for storing an estimation model configured to output, in response to input of one or more items of patient information, a candidate disease that a person related to the patient information may be suffering from among a plurality of diseases; The processor, Obtaining input of values ​​of one or more items of patient information related to the patient; By inputting the values ​​of the one or more items into the estimation model, candidates of diseases that the patient may be suffering from are obtained; Obtaining additional required information to be obtained from the patient based on the disease candidate; The communication interface presents the additional request information.