JP6844088B2 - Medical information provision method, medical information provision device, and program - Google Patents

Description

The present invention relates to a medical information providing method, a medical information providing device, and a program for providing medical information such as a possible disease to a user.

A server installed in a service institution that provides medical information stores a pathological condition database that stores the types of symptoms that occur in each part of the body, the names of injuries and illnesses estimated for each degree, and their probabilities. However, when the terminal is used to connect to the pathological condition database in the server and the applicable symptom is selected, the name of the injury or illness estimated by the server is searched, and the pathological condition estimation system in which the result is displayed on the terminal is known. (For example, Patent Document 1).

Japanese Unexamined Patent Publication No. 2004-240734

However, in such a pathological condition estimation system, it is difficult to efficiently narrow down possible diseases because the user only inputs a symptom and the system does not ask a question to the user.

Therefore, an object of the present invention is to provide a medical information providing method, a medical information providing device, and a program capable of efficiently narrowing down possible diseases.

In order to solve the above-mentioned problems, in the first aspect of the present invention, a specific symptom is selected from a plurality of symptoms according to a user input, a first numerical value indicating the frequency of occurrence of each of the plurality of diseases, and a first numerical value indicating the frequency of occurrence of each of the plurality of diseases. The morbidity probability for each of the plurality of diseases is calculated based on the second numerical value indicating the frequency of occurrence of the specific symptom in each of the plurality of diseases, and corresponds to the morbidity probability which is equal to or higher than a preset threshold value. A question to ask the details of the specific symptom associated with the specific symptom and options for the question, and the presence or absence of a predetermined number of the plurality of symptoms other than the specific symptom. Provided is a method for providing medical information, characterized in that at least one of them is determined as a question to be presented to a user together with the selected disease.
Here, the morbidity probability refers to the degree to which the input symptom group is medically closer to the symptom group of which disease.

In the medical information providing method of the present invention having the above configuration, when there is the medical inquiry that the user has not answered, the medical inquiry and the corresponding option are output, and the user has answered all of the medical questions. In the case of, it is preferable to output the presence or absence of the predetermined number of symptoms.

Further, in the medical information providing method of the present invention having the above-described configuration, among the symptoms that can occur in the disease that can cause the specific symptom, the option of the disease that can cause the specific symptom depending on the presence or absence of the symptom is selected. It is preferable to select the presence or absence of the predetermined number of symptoms so as to reduce the number, and to determine the question based on the result of the selection.

Further, in the medical information providing method of the present invention having the above-described configuration, among the symptoms that can occur in the diseases that can cause the specific symptoms, the number of diseases that can cause the specific symptoms depends on the presence or absence of the symptoms. It is preferable to select the question of the presence or absence of the predetermined number of symptoms so as to approach equal division.

Further, in the medical information providing method of the present invention having the above-described configuration, it is preferable to calculate the morbidity probability based on a third numerical value indicating the rate of input errors by the user.

Further, in the medical information providing method of the present invention having the above-described configuration, a fourth method of acquiring data indicating gender in response to the user input and indicating the frequency of occurrence of each of the plurality of diseases in the gender. It is preferable to calculate the morbidity probability based on the numerical value.

Further, in the medical information providing method of the present invention having the above-described configuration, data indicating an age group is acquired in response to the user input, and the frequency of occurrence of each of the plurality of diseases in the age group is indicated. It is preferable to calculate the morbidity probability based on the numerical value of 5.

Next, in the second aspect of the present invention, a selection unit that selects a specific symptom from a plurality of symptoms according to user input, a first numerical value indicating the frequency of occurrence of each of the plurality of diseases, and the plurality of the above. Based on the second numerical value indicating the frequency of occurrence of the specific symptom in each of the diseases, the morbidity probability is calculated for each of the plurality of diseases, and the disease corresponding to the morbidity probability which is equal to or higher than a preset threshold value. The calculation unit that selects the specific symptom, the interview asking the details of the specific symptom associated with the specific symptom, the options for the interview, and the presence or absence of a predetermined number of the plurality of symptoms other than the specific symptom. Provided is a medical information providing device comprising at least one of the above, a display control unit for determining a question to be presented to a user together with a disease selected by the calculation unit, and a display control unit.

In the medical information providing device of the present invention having the above configuration, when the user has not answered the inquiry, the display control unit outputs the inquiry and the corresponding option, and the user outputs the inquiry and the corresponding option. When all the interviews have been answered, it is preferable to output the presence or absence of the predetermined number of symptoms.

Further, in the medical information providing device of the present invention having the above-described configuration, among the symptoms that can occur in the disease that can cause the specific symptom, the option of the disease that can cause the specific symptom depending on the presence or absence of the symptom is selected. Further having a symptom presence / absence selection unit for selecting the presence / absence of the predetermined number of symptoms so as to reduce the number, the display control unit determines the question based on the selection result of the symptom presence / absence selection unit. Is preferable.

Further, in the medical information providing device of the present invention having the above-described configuration, the calculation unit may calculate the morbidity probability based on a third numerical value indicating the rate of input errors by the user. preferable.

Further, in the medical information providing device of the present invention having the above-described configuration, the selection unit can acquire data indicating gender in response to the user input.
It is preferable that the calculation unit further calculates the morbidity probability based on a fourth numerical value indicating the frequency of occurrence of each of the plurality of diseases in the sex.

Further, in the medical information providing device of the present invention having the above-described configuration, the selection unit acquires data indicating an age group in response to the user input, and the calculation unit obtains the plurality of data indicating the age group. It is preferable to calculate the morbidity probability based on a fifth numerical value indicating the frequency of occurrence of each of the diseases.

Further, a third aspect of the present invention includes a procedure for selecting a specific symptom from a plurality of symptoms in response to a user input to a medical information providing device that provides medical information to a user, and a procedure for selecting a specific symptom from a plurality of diseases, respectively. A procedure for calculating the morbidity probability for each of the plurality of diseases based on a first numerical value indicating the frequency of occurrence of the disease and a second numerical value indicating the frequency of occurrence of the specific symptom in each of the plurality of diseases. A procedure for selecting a disease corresponding to the prevalence of the disease, which is equal to or higher than a preset threshold, an interview asking the details of the specific symptom associated with the specific symptom, options for the interview, and the specific symptom. Provided is a procedure for determining at least one of a predetermined number of symptoms among the plurality of symptoms other than the above as a question to be presented to a user together with the selected disease, and a program for executing the procedure.

According to the present invention, it is possible to provide a medical information providing method, a medical information providing device, and a program capable of efficiently narrowing down possible diseases.

It is a figure which shows the whole structure of the medical information provision system 1 which concerns on embodiment of this invention. It is the schematic which shows the structure of the server 10. It is a figure which illustrates the data structure of a dictionary master. It is a figure which illustrates the data structure of the symptom master. It is a figure which illustrates the data structure of the additional interview master. It is a figure which illustrates the data structure of the disease master. It is a flowchart which shows the operation of a server. It is a figure which shows an example of the initial screen displayed on a user terminal. It is a figure which shows an example of the screen of the user terminal which displayed the input data and the list of candidate symptoms. It is a figure which shows an example of the screen of the user terminal which displayed the list of possible illnesses, and the additional inquiry. It is a figure which shows an example of the screen of the user terminal which displayed the list of possible diseases and the question whether or not there is a symptom.

Hereinafter, typical embodiments of the present invention will be described in detail with reference to the drawings. Here, a medical information providing system including a server as a medical information providing device and a user terminal and a medical information providing method using the same will be given as an example, but the present invention is not limited to these drawings. Since the drawings are for conceptually explaining the present invention, the dimensions, ratios, or numbers may be exaggerated or simplified as necessary for easy understanding.

[Overall system configuration]
FIG. 1 shows the overall configuration of the medical information providing system 1. The medical information providing system 1 is an information system that provides a user with medical information including a disease that may be affected. As shown in FIG. 1, a user connected to the server 10 and the server 10 via a communication network. The terminal 20 and the like are included in the configuration. Here, the user is a patient (including a potential patient), and the medical information providing system 1 can be used by the patient to know the disease that he / she may be suffering from. .. The medical information providing system 1 may be used by a doctor.

The server 10 is a computer that processes medical information according to the operation of the user terminal 20, and is an example of a medical information providing device. The configuration of the server 10 will be described later. The server 10 may be one or a plurality of servers.

The user terminal 20 is a computer used by the user, and includes, for example, a desktop computer, a notebook computer, a tablet terminal, a smartphone, and a mobile phone. In FIG. 2, one desktop computer 21 and one smartphone 22 are shown as an example of the user terminals 20, but a larger number of user terminals may be connected to the network. The user terminal 20 has an input means for the user to input data such as a symptom, a communication means for communicating information between the server 10 by wire or wirelessly, and a display screen for displaying received medical information. And have at least.

In such a medical information providing system 1, when a user inputs data such as symptoms via a user terminal 20, the server (medical information providing device) 10 calculates the morbidity probability for each disease based on the input data. , The disease name which is the calculation result is output to the user terminal 20. By displaying the disease name on the user terminal 20, the user can know the possible diseases. Here, the morbidity probability refers to the degree to which the input symptom group is medically closer to the symptom group of which disease.

Further, the server (medical information device) 10 may output various medical information such as the degree of possibility of illness, the need for emergency response, and the clinical departments and medical institutions that can respond, together with the disease name. Good. As a result, the user can obtain comprehensive medical information including the disease and its coping method.

[Server configuration]
FIG. 2 schematically shows the configuration of the server 10. As shown in FIG. 2, the server 10 includes a CPU 11, a RAM 12, a ROM 13, and a communication interface 14. Each of these elements is communicably connected to each other via the bus 15. Hereinafter, each element will be described in order.

The CPU 11 is a device that performs various calculations, and particularly executes functions as a search unit 111, a data selection unit 112, a disease probability calculation unit 113, a symptom presence / absence question selection unit 114, and a display control unit 115. Such a description of each functional unit will be given after the description of the ROM 13.

The RAM 12 is a main storage device (memory). The CPU 11 loads the program and data (for example, precompiled data for probability calculation) necessary for realizing each of the above functional units into the RAM 12, performs arithmetic processing, and then returns the program to the RAM 12.

The communication interface 14 is an interface for communicating with the user terminal 20 and other servers. For example, the operation information is received from the user terminal 20, and the calculation result of the CPU 11 is transmitted to the user terminal 20.

The ROM 13 is an auxiliary storage device, and stores the dictionary master 131, the symptom master 132, the additional inquiry master 133, and the disease master 134 as described in detail below. The ROM 13 also stores the input data received by the communication interface 14 and the program required to realize each functional unit of the CPU 11.

(Dictionary master)
FIG. 3 illustrates the data structure of the dictionary master 131. As shown in FIG. 3, the dictionary master 131 stores IDs, formal expressions, and variations such as notational fluctuations and synonyms for each of the plurality of symptoms. For example, in addition to the formal notations such as "headache" and "nausea", variations such as "headache", "thrilling", "hakike", "nausea", and "nausea" are available. ..

By preparing formal expressions and variations in this way, when all or part of the input keywords correspond to the expressions of variations, a list of candidates for formal notation is presented to the user, for example, as shown in FIG. , Allows the user to select. This is because the keywords are not always entered in a certain form (formal expression) such as "headache" and "nausea", but various such as "headache" and "likely to vomit from yesterday". It is considered that it can take a different shape.

(Symptom master)
FIG. 4 illustrates the data structure of the symptom master 132. As shown in FIG. 4, the symptom master 132 includes _{numerical data aij} indicating how often the symptom j appears in the disease i for each of the plurality of diseases. Such numerical data _aij is set based on medical data such as a treatise and the knowledge of a doctor, and is used for calculating the disease probability described later. The numerical value _aij once set can be updated according to the output of the medical information providing device, thereby improving the accuracy of the medical information provided to the user.

The symptom master 132 includes a symptom ID and a non-symptom ID for each symptom. Here, the complaint symptom ID is an ID indicating that there is a specific symptom, and is output when the symptom is input as "yes" by the user. The non-complaint symptom ID is an ID indicating that there is no specific symptom, and is output when the symptom is input as "none" by the user. For example, in FIG. 4, the symptom ID “A1” and the non-complaint symptom ID “B1” are assigned to the symptom “with fever” and “without fever”. The symptom ID and the non-symptom ID may be collectively referred to as a symptom ID.

The symptom master 132 may include numerical data regarding the gender dependence of the symptom. For example, by retaining data that men do not have menstrual pain, it is possible to prevent the display of questions such as "Do you have menstrual pain?" After selecting a man.

(Additional interview master)
FIG. 5 illustrates the data structure of the additional interview master 133. As shown in FIG. 5, the additional interview master 133 includes additional interviews asking for more detailed conditions regarding some of the symptoms registered in the symptom master 131, and options for answering the corresponding additional interviews k. , Are set individually. For example, "How long has it been since you coughed?" Is set as one of the additional interviews for the symptom "cough", and "within 3 days" and "within 1 month" as options for answering the additional interview. "And" more "are provided.

The additional interview master 133 also holds _{numerical data bik} indicating the evaluation of the validity of the disease i for each option k of the answer to the additional interview for each of the diseases in which a predetermined symptom occurs. Numerical data _bik is, for example, when the answer "within 3 days" (answer ID "Z1") is given to the additional question "How long has it been since the cough started?" For the symptom "cough". , The degree to which the disease "cold (acute upper respiratory tract inflammation)", "bronchial asthma", "whooping cough" ... becomes more probable, or conversely, the probability is reduced.

In this way, when a specific symptom is input, an additional interview is conducted and an answer is obtained from the user, so that a more detailed symptom can be obtained. As will be described later, by reflecting the detailed symptoms obtained from the answers to the additional interviews in the calculation of the disease probability, the accuracy of the output of the medical information providing device is improved. In addition, it is possible to efficiently narrow down possible diseases with a smaller number of user inputs.

(Disease master)
FIG. 6 illustrates the data structure of the disease master 134. Disease master 134, as shown in FIG. 6, for each registered disease, numerical data c _i that indicates the frequency of occurrence of the disease i in the world has been defined. Numerical data c _i is set based on medical data and knowledge of the physician, such as paper, used to calculate the probability of disease to be described later. Numerical data c _i is capable appropriately updated, thereby to improve the accuracy of the medical information provided to the user.

The disease master 134 also defines _{numerical data dil} indicating the susceptibility (gender dependence) of the disease i according to the sex l for each of the registered diseases. The numerical data _il is set based on medical data such as a treatise and the knowledge of a doctor, and is used for modifying the disease probability described later.

The disease master 134 further defines _{numerical data im} indicating the susceptibility of the disease i according to the age group m for each of the registered diseases. For example, the age group m is divided into "less than 1 month old", "under 1 year old", "1 year old and under 4 years old", "4 years old and under 10 years old", "10 years old and under 20 years old", "20 to 30 years old", "30 to 40 years old", "40 to 50 years old", "50 to 60 years old", "60 to 70 years old", "70 to 80 years old" There are 12 levels of "under the age" and "over 80". In addition, the numerical data _im is set based on medical data such as a treatise and the knowledge of a doctor, and is used for modifying the disease probability described later.

For each registered disease, the disease master 134 includes a commentary on the disease, possible symptoms, causes, treatment methods, therapeutic agents, necessity of emergency response, precautions, corresponding clinical departments, and available medical care. It may include disease information such as an institution. By providing such disease information to the user together with the name of a possible disease, it is possible to provide the user with useful medical information.

The symptom master 132, the additional interview master 133, and the disease master 134 described above may be processed to improve the memory utilization efficiency and the processing efficiency, and may be loaded into the memory as precompiled data (see FIG. 7).

(Search department)
The description of the functional part of the CPU 11 will be started.
The search unit 111 searches the dictionary master 131 when a keyword is input by the user. Then, when the keyword corresponds to any of the formal expressions registered in the dictionary master 131, the search unit 111 outputs the complaint symptom ID corresponding to the corresponding symptom.

If the input keyword does not correspond to any of the formal expressions, the search unit 111 searches whether all or part of the keyword corresponds to the variation registered in the dictionary master 131. Then, the search unit 111 outputs a formal expression of all the corresponding symptoms to the user terminal 20 via the communication interface 14. When the user selects one of the symptoms from the list of formal expressions displayed on the user terminal 20 (see FIG. 9), the corresponding complaint symptom ID is output from the search unit 111 (or data selection unit 112). The symptom.

(Data selection section)
The data selection unit 112 acquires the data selected by the user, passes it to the disease probability calculation unit 113, and stores it in the ROM 13.

(Disease probability calculation department)
The disease probability calculation unit 113 calculates the morbidity probability for each disease based on the above-mentioned input data, and extracts possible diseases. That is, the disease probability calculation unit 113, with respect to frequency (initial value) _{c i} of each disease, numerical _a ij based on the input _data, b _ik, d _il, by multiplying the _{e im,} morbidity per disease Repeat the process of calculating the possibility (posterior probability).

<Disease probability calculation method>
An example of a method for calculating the susceptibility (posterior probability) for each disease will be specifically described.
Here, the posterior probability "postP _i " is set as a relative value indicating the probability of the disease i in consideration of the fact that certain information is input, and the prior probability "preP _i " is set as the relative value before the input of certain information. It is a relative value indicating the probability of disease i. In this embodiment, since the variables of the symptom j, the answer k of the additional interview, the gender l, and the age group m are used, the posterior probability of the disease i is expressed as postP _ijklm and the prior probability is expressed _{as preP ijklm.} However, in order to avoid complication, variables that do not appear in the formula are omitted in the following formula.

When the disease probability calculation unit 113 calculates the disease probability based on the symptom ID, the method for calculating the disease probability differs depending on whether the selected data is the complained symptom ID or the non-complaint symptom ID.

The disease probability calculation unit 113 receives the complaint symptom ID when a specific symptom such as "fever" is input as "present". Then, the disease probability calculation unit 113, and numerical a _ij indicating the frequency with which heat is out in each disease elected, by taking the product of the prior probability c _i, each possible disease when there is a particular condition ( The posterior probability) is calculated.

Expressed this process in a formula, the disease probability calculation unit 113, the chromatic訴症shaped j is input, using the following formula 1, including the defined variables a _ij Symptom master 132, the posterior of each disease i to calculate the probability postP _ij.
postP _ij = preP _ij * a _ij ... (Equation 1)
As described above, the variable _{c i} is used as the initial value of PReP _ij.
However, since the complaint symptom j is not limited to one and a plurality of symptom j can be selected, the disease probability calculation unit 113 repeats the calculation by the equation 1 for the number of the selected symptom. At this time, the postP _{ij of the formula 1 or the postP ij of the} following formula 2 is substituted into _{the preP ij of the} second and subsequent orders according to the previously input symptom ID.

Further, the disease probability calculation unit 113 receives the non-complaint symptom ID when a specific symptom such as "fever" is input as "none". In this case, the disease probability calculation unit 113 calculates the possibility of each disease by performing the same processing as when the complaint symptom ID is received. However, at this time, the value to be producted with the prior probability is (1-a _ij ) _{, not the numerical value a ij.} For example, if the frequency _a 11 = 0.7 exiting symptoms "fever" in diseases 4 "cold", here will be multiplied by 1-0.7 = 0.3.

Expressed this process in a formula, the disease probability calculation unit 113, when no訴症like is input, the using the following formula 2, the posterior probability PostP _ij for each disease i 'containing _{(1-a ij)} calculate.
postP _ij '= preP _ij * (1-a _ij ) ・ ・ ・ (Equation 2)
Here, the postP _{ij of the formula 1 or the postP ij of the} formula 2 is substituted into _{the preP ij of the} formula 2 according to the symptom ID input last time.

Then, if the answer to additional inquiry is inputted, the disease probability calculation unit 113, by calculating the posterior probability PostP _ik using the formula 3 containing variables b _ik defined additional inquiry master 133, disease Calculate (modify) the probability.
_{postP ik = preP ik * b ik} ··· ( Equation 3)
Here, the PReP _ik of formula 3, so that the PostP _ij of PostP _ij or formula 2 of the formula 1 'is substituted.

Also, if the sex is entered, the disease probability calculation unit 113, by calculating the posterior probability PostP _il using the formula 4 including the variable d _il defined disease master 134, calculates the disease probability ( Qualify.
postP _il = preP _il * d _il ... (Equation 4)
Here, the PReP _il of formula 4, postP _ij of formula 1, postP _ij of formula 2 ', or so that the PostP _ik of formula 3 is substituted.

Also, if the age is entered, the disease probability calculation unit 113, by calculating the posterior probability PostP _im using Equation 5 follows, including a defined variable e _im disease master 134, calculates the disease probability (Qualify.
postP _im = preP _im * e _im ... (Equation 5)
Here, the PReP _im of formula 5, comprising PostP _ij of formula 1, postP _ij of formula 2 ', postP _ik of formula 3, or to PostP _il of formula 4 are substituted.

In this way, the disease probability calculation unit 113 calculates the posterior probability PostP _i for each disease i registered in the condition master 132. Then, the disease probability calculation unit 113 sorts the diseases in descending order of posterior probabilities, makes them absolute by taking the reciprocal of the sum of all relative probabilities, and extracts only the diseases having a value equal to or higher than the threshold value. Alternatively, it may be said that the top diseases are selected.

<Consideration of the possibility of input error>
The disease probability calculation unit 113 considers the possibility of an input error when calculating the posterior probability described above. Here, the input error includes both an error in which the user inputs something as "not" and an error in which the user inputs something as "yes".

To give an example of an input error, when there is a heat generation of 37.0 degrees, the user thinks that "this is not heat generation" and inputs "no heat generation". Also, if a symptom that did not occur on that day appears on the next day, the user will say, "The symptom has not yet appeared on the previous day, even though the disease is the same on that day and the next day. It is possible to enter "no symptom".

In these two examples, the situation in which the user inputs "no symptom" can be fully understood, but the accuracy of the calculation of the morbidity probability is improved when the user inputs "with symptom". This is not limited to the problem of high accuracy and low accuracy, but it is better not to enter anything than to enter the original symptom as "no symptom" (that is, the more you enter, the worse the accuracy becomes. It also means that a situation such as) will occur.

Therefore, in the present embodiment, the possibility of an input error is taken into consideration when calculating the disease probability to provide elasticity to absorb the above-mentioned problem.

An example of a specific method is given.
An error in which the user inputs something as "not", that is, an error in which the original symptom of the disease is said to be "not", can occur at a reasonable frequency. Therefore, in the present embodiment, the user makes this kind of mistake at a predetermined frequency (for example, 20%), and sets a value α indicating this.

Also, it is less likely that the user will say "yes" what they don't have. In this case, it is not necessary to consider the possibility that the symptom has not yet appeared, and there is a high possibility of a pure user misunderstanding (for example, a mistake in clicking a button or a misinterpretation of a medical term). it is conceivable that. Therefore, in the present embodiment, it is assumed that the user makes this kind of mistake at a certain frequency (for example, 5%), and a value β indicating this is set.

Then, the following equation 6 including the product of the right side of the above equation 1 and the value (1-α) indicating the ratio at which the input is accurately performed and the product of the right side of the above equation 2 and the value β is used. By calculating the posterior probability, the possibility of this kind of error is taken into consideration.
postP _ij = preP _ij * a _ij * (1-α)
+ PreP _ij * (1-a _ij ) * β ・ ・ ・ (Equation 6)

Further, similarly to the above equation 6, the posterior probability is calculated by using the product of the right side of the equation 2 and the value (1-β) and the product of the right side of the equation 1 and the value α.
postP _ij '= preP _ij * (1-a _ij ) * (1-β)
＋ preP _ij * a _ij * α

In this way, it is possible to create a formula for calculating the morbidity probability in consideration of an input error regarding the presence or absence of symptoms. Similarly, a formula for calculating the probability of morbidity can be created that takes into account input errors regarding the answers to additional interviews, gender, and age group.
By using such a calculation formula, it is possible to prevent possible diseases from being excluded or ranked low due to a user input error. Therefore, the reliability of the calculation result is improved.

(Question selection department for the presence or absence of symptoms)
The symptom presence / absence question selection unit 114 selects a symptom presence / absence question that should be displayed together with a possible disease. Specifically, the symptom presence / absence question selection unit 114 selects a predetermined number (for example, three) of symptom presence / absence that minimizes the choice of diseases that can cause the symptom selected by the user. This works in the direction of selecting a predetermined number of symptoms that minimizes information entropy (sometimes referred to as the average amount of information; the same applies hereinafter).

A specific example will be described.
When the user selects "cough = yes", it is assumed that there are 100 registered diseases that cause coughing. Further, it is assumed that the total number of symptoms that can occur other than cough in the 100 diseases is 250 (for example, in influenza, many symptoms such as fever, arthralgia, and chills occur in addition to cough).

It is considered that these 250 symptoms include rare symptoms that appear only in some diseases. Questions asking for the presence or absence of these unusual symptoms are less efficient than others because it is more difficult to reduce information entropy.

Therefore, with respect to the 250 symptoms that can occur in the above-mentioned 100 diseases, the symptoms that appear in 50 diseases but not in the remaining 50 diseases (that is, the symptoms that appear or do not appear at a ratio of 50:50). ), Then the presence or absence of symptoms that appear or do not appear at a ratio of 49:51 or 51:49, and so on. , It can be said that this is a question that makes it easy to reduce information entropy.

Then, for example, the following equation 8 always holds.
50 * 50> 51 * 49> 52 * 48> ・ ・ ・> 99 * 1> 100 * 0 ・ ・ ・ (Equation 8)
Considering this, let n be the number of symptom j (among N diseases to which symptom j corresponds) that appears in a specific disease i out of N (for example, 100 diseases described above). For 0 ≦ n ≦ N), the value given by the following equation 9 is calculated.
n * (Nn) ... (Equation 9)
When a predetermined number of symptoms j having the maximum value are selected, these symptoms minimize the information entropy.

A question regarding the presence or absence of the symptom selected in this way is presented to the user, and the user answers the question, so that the possible diseases can be efficiently narrowed down.

(Display control unit)
The display control unit 115 determines an additional question to be displayed on the user terminal together with the calculation result (possible disease) of the disease probability calculation unit 113. Specifically, the display control unit 115 displays such unanswered additional interviews when there are unanswered additional interviews related to the symptom selected by the user. If there are no unanswered additional interviews, the display control unit 115 determines the presence or absence of a predetermined number of symptoms based on the selection result of the symptom presence / absence question selection unit 114.

[Server operation]
The operation of the server 10 having the above-described configuration and the procedure for providing medical information will be described with reference to FIGS. 7 to 11. FIG. 7 is a flowchart showing the operation of the server 10. 8 to 11 are examples of screens displayed on the user terminal 20.

The procedure for providing medical information is roughly as follows: (1) search and data selection associated with user operations in steps S1 and S2, (2) calculation of disease probability in step S3, and (3) selection of symptom presence / absence questions in step S4. , And (4) presentation of the calculation result to the user in steps S5 to S8 are repeated. These procedures will be described in detail below.

(Search and data selection associated with user operations)
When the procedure starts, the initial screen is displayed on the user terminal 20. The initial screen prompts the input of a keyword in the input field 31, as shown in FIG. 8, for example. On the initial screen, along with the input field, the most searched symptom within a certain period may be displayed for the user's selection.

When the user inputs a keyword, the search unit 111 searches the dictionary master 131 in step S1. As described above, the search unit 111 outputs the complained symptom ID of the corresponding symptom when the keyword corresponds to the formal expression registered in the dictionary master 131.

The search unit 111 also outputs a list of formal expressions (and symptom IDs) of the corresponding symptom when all or part of the input keyword corresponds to the variation registered in the dictionary master 131. .. As a result, the user terminal 20 displays a list of candidates 32 as shown in FIG. 9, for example. When the user selects one of the listed candidates, the corresponding complaint symptom ID is output.

Further, when it is determined in step S9, which will be described later, that the user operation is continuing, in step 1, the search unit 111 (or the data selection unit 112) determines the complaint symptom ID and the non-complaint symptom according to the user operation. The ID, the answer ID to the additional interview, the age group, and the gender information are output.

Next, in step 2, the input data (complaint symptom ID, non-complaint symptom ID, response ID to the additional inquiry, age group data, and gender data) output in step S1 are stored in the ROM 13.

Prior to steps S1 and S2, the data related to the symptom master 132, the additional interview master 133, and the disease master 134 are processed in order to improve the memory utilization efficiency and processing efficiency at the time of calculation (data prescription in FIG. 7). processing). The preprocessed data in this way is loaded into memory as precompiled data.

(Calculation of disease probability)
The procedure for calculating the disease probability in step S3 involves calculating the disease probability from the symptom ID in step S31, and modifying the probability distribution based on the additional interview data, gender data, and age group data in steps S32 to S34, and then step S35. Probability descending order sorting and selection of top diseases in. Hereinafter, steps S31 to S35 will be described in order.

In step S31, the disease probability calculation unit 113 calculates the disease probability based on the symptom ID. Here, in order to consider the possibility of an input error as described above, Equations 6 and 7 are used instead of Equations 1 and 2 described above. As a result, even if the frequency of occurrence of the symptom is defined as zero (for example, the frequency of occurrence of the symptom "fever" for "myopia", the initial value is multiplied by zero to obtain the posterior probability of the disease. It never goes to zero, which means that it does not completely rule out the possibility of a disease.

In step S32, the disease probability calculation unit 113 calculates (modifies) the disease probability based on the answers to the additional interviews. As a result, the disease probability can be calculated based on more detailed symptoms, so that a more reasonable calculation result can be provided to the user.

In step S33, the disease probability calculation unit 113 calculates (modifies) the disease probability based on the gender selected by the user. Further, in step S34, the disease probability calculation unit 113 calculates (modifies) the disease probability based on the age group selected by the user. By considering the possibility of the disease based on more personal information such as gender and age group, the possible diseases can be evaluated more individually.

By going through such steps S31 to S34, it becomes possible to calculate the disease probability (posterior probability) based on the deeper symptom and more individual information registration in addition to the user's input, and as a result, the disease probability (posterior probability) can be calculated. The medical information provided to the user becomes more relevant.

Moreover, in the present embodiment, the possibility of a user input error is taken into consideration when calculating the disease probability. In addition, in the present embodiment, instead of extracting and sorting only the diseases that cause the symptoms selected by the user, the possibility of all the registered diseases is always considered. This makes it possible to effectively prevent the possible diseases from being leaked from the calculation results.

Then, in step S35, the disease probability calculation unit 113 extracts only the higher-ranking diseases that exceed a preset threshold value (for example, 1%) from all the calculated posterior probabilities, and displays them on the user terminal 20 in descending order. Output to make it.

(Selection of questions regarding the presence or absence of symptoms)
In step S4, the symptom presence / absence question is selected in the symptom presence / absence question selection unit. Step S4 is a step of extracting a predetermined number (three in this case) of questions to be answered with “yes” and “no” as shown in FIG. 11, for example, calculating the discrimination ability score in step S41 and step S42. Includes selection of symptoms with the highest discrimination score. Here, the discrimination ability score is a value calculated using the above-mentioned equation 9. In this embodiment, the presence or absence of the top three symptoms of the discrimination ability score is asked.

That is, in step S41, the symptom presence / absence question selection unit 114 calculates the discrimination ability score for all the symptoms included in the disease that can cause the symptom selected by the user. Next, in step S42, the symptom presence / absence question selection unit 114 selects the top three symptom of the discrimination ability score.

By selecting the presence or absence of a predetermined number of symptoms in this way, it is possible to actively ask the user to narrow down the possible diseases with a smaller number of data selections.

(Presentation of calculation result to user)
In step S5, the calculation result (possible disease) output in step 3 is output to the user terminal 20. As a result, the user terminal 20 displays the names of possible diseases, for example, as shown in FIGS. 10 and 11. At this time, for example, disease information such as the degree of possibility of the disease (including display by numerical values and graphs), definition of the disease, urgency of treatment, corresponding clinical department, and available medical institution may be displayed. ..

At the same time, in step S6, the display control unit 115 confirms whether or not there is an unanswered additional inquiry related to the symptom selected by the user. When there is an unanswered additional inquiry, the display control unit 115 instructs the user terminal 20 to display the additional inquiry in step S7. As a result, when the user selects a specific symptom registered in the additional interview master 133, several "additional interviews" are continuously displayed on the subsequent screens. For example, in FIG. 10, when the symptom "nausea" is selected, questions such as "when did you start nausea?" And "did you eat raw food?" Are continuously displayed on the subsequent screens.

The additional interviews are registered in the additional interview master based on the thinking process of the doctor. Therefore, for the doctor who is the user, the medical information providing system 1 can be used as confirmation of the contents of the inquiry to the patient. Further, for the patient who is a user, useful medical information can be obtained by responding to the medical information providing system 1 as if he / she is being examined by a doctor.

Further, when all the additional interviews have been answered, the display control unit 115 instructs the user terminal 20 to display the question of the presence or absence of the symptom in step S8 according to the selection result in step S4. As a result, the presence or absence of a predetermined number (for example, three) of symptoms will be displayed on the user terminal. For example, in FIG. 11, the presence or absence of three symptoms such as weakness, dizziness, and lightheadedness is asked for the selected symptom “palpitations”.

Finally, in step S9, it is confirmed whether or not the user operation is continuing. When the continuation is confirmed, the steps S1 to S8 are repeated, and if the continuation is not confirmed, the procedure ends.

Although the typical embodiments of the present invention have been described above, the present invention is not limited to these, and various design changes are possible, and these are also included in the present invention.
For example, the user terminal itself may also have a function as the medical information providing device 1.
Further, the dictionary master, the symptom master, the additional inquiry master, and the disease master do not need to be classified in this way, and may be provided as, for example, an integrated database, or may be further divided.

1 ... Medical information provision system,
10 ... server,
20, 21, 22, ... User terminal,
111 ... Search section,
112 ... Data selection unit,
113 ... Disease probability calculation unit,
114 ... Symptoms Question Selection Department,
115 ... Display control unit,
131 ... Dictionary master,
132 ... Symptom master,
133 ... Additional interview master,
134 ... Disease master.

Claims (7)

The computer
Select a specific symptom from multiple symptoms according to user input,
Based on the first numerical value indicating the frequency of occurrence of each of the plurality of diseases and the second numerical value indicating the frequency of occurrence of the specific symptom in each of the plurality of diseases, the probability of morbidity for each of the plurality of diseases is determined. Calculate and
Select a disease corresponding to the morbidity probability that is equal to or higher than a preset threshold value.
At least one of a question about the details of the particular symptom associated with the particular symptom, options for the question, and the presence or absence of a predetermined number of the plurality of symptoms other than the particular symptom. determined as a question to be presented to the user together with the selected the disease,
In order to reduce the choice of diseases that can cause the specific symptoms, the number of diseases that can cause the specific symptoms is divided equally among the symptoms that can occur in the diseases that can cause the specific symptoms, depending on the presence or absence of the symptoms. Select the question of the presence or absence of the predetermined number of symptoms so as to approach, and determine the question based on the result of the selection.
A method of providing medical information characterized by this. If there is the question that the user has not answered, the question and the corresponding option are output.
When the user has answered all of the interviews related to the selected symptom among the interviews, the question of the presence or absence of the predetermined number of symptoms is output.
The medical information providing method according to claim 1. To calculate the morbidity probability based on a third numerical value indicating the rate of input errors by the user.
The medical information providing method according to claim 1 or 2. In response to the user input, data indicating gender is acquired, and
To calculate the prevalence probability based on a fourth numerical value indicating the frequency of occurrence of each of the plurality of diseases in the sex.
The medical information providing method according to any one of claims 1 to 3, wherein the medical information is provided. In response to the user input, data indicating the age group is acquired, and
To calculate the prevalence probability based on a fifth numerical value indicating the frequency of occurrence of each of the plurality of diseases in the age group.
The medical information providing method according to any one of claims 1 to 4, wherein the medical information is provided. A selection unit that selects a specific symptom from multiple symptoms according to user input,
Based on the first numerical value indicating the frequency of occurrence of each of the plurality of diseases and the second numerical value indicating the frequency of occurrence of the specific symptom in each of the plurality of diseases, the probability of morbidity for each of the plurality of diseases is determined. A calculation unit that calculates and selects a disease corresponding to the symptom probability that is equal to or higher than a preset threshold value.
At least one of a question about the details of the particular symptom associated with the particular symptom, options for the question, and the presence or absence of a predetermined number of the plurality of symptoms other than the particular symptom. It is provided with a display control unit that determines a question to be presented to the user together with the disease selected by the calculation unit.
The predetermined number of symptoms so that the display control unit approaches the number of diseases that can cause the specific symptom evenly depending on the presence or absence of the symptom from among the symptom that can occur in the disease that can cause the specific symptom. Select the question of presence or absence, and determine the question based on the result of the selection.
Medical information providing apparatus according to claim and this. For medical information providing devices that provide medical information to users
The procedure for selecting a specific symptom from multiple symptoms according to user input,
Based on the first numerical value indicating the frequency of occurrence of each of the plurality of diseases and the second numerical value indicating the frequency of occurrence of the specific symptom in each of the plurality of diseases, the probability of morbidity for each of the plurality of diseases is determined. Procedure to calculate and
A procedure for selecting a disease corresponding to the prevalence probability, which is equal to or higher than a preset threshold value, and
At least one of a question about the details of the particular symptom associated with the particular symptom, options for the question, and the presence or absence of a predetermined number of the plurality of symptoms other than the particular symptom. The procedure for determining the question to be presented to the user together with the selected disease, and
From the symptoms that can occur in the diseases that can cause the specific symptoms, the question of the presence or absence of the predetermined number of symptoms is selected so that the number of diseases that can cause the specific symptoms approaches evenly depending on the presence or absence of the symptoms. Then, the procedure for determining the question based on the result of the selection and
A program to execute.

Images