JPWO2017183085A1 - Interview support system - Google Patents

Abstract

The inquiry support system 10 includes a reception unit 26 that receives user input, an inquiry holding unit 28 that holds a plurality of inquiry items, and an inquiry item based on at least one of the first function and the second function. An output determination unit that determines whether or not there is an output, a sending unit 44 that outputs an inquiry item based on the determination of the output determination unit, and first evaluation information for the inquiry item that is output based on the first function The first evaluation input unit 43 and a second evaluation input unit 45 for receiving second evaluation information for the inquiry items output based on the second function. The output determination unit adjusts a value for determining the presence or absence of the output of the inquiry items or the priority of the output based on the evaluation information of at least one of the first evaluation information and the second evaluation information.

Description

  The present invention relates to an inquiry support system that outputs an inquiry item for diagnosing a patient.

  When diagnosing the examinee, the doctor interrogates the examiner and derives an appropriate disease name while eliminating the possibility of various diseases. When a doctor diagnoses a large number of examinees within a limited consultation time at a medical site, the time available for one examinee is limited, for example, about 3 minutes for one examinee. May be diagnosed with Therefore, a system that supports an inquiry so that a doctor can efficiently diagnose and check the state is desired.

  Patent Document 1 discloses a medical interview system for a medical examinee having a display device for displaying a common inquiry and an inquiry question specific to each department and an expected answer corresponding thereto. In this medical examination system for medical examinees, a common inquiry and an inquiry question specific to each department are displayed according to a predetermined order.

JP 2005-321472 A

  In the technique disclosed in Japanese Patent Laid-Open No. 2004-228688, the question for inquiry specific to each department is displayed according to a predetermined order. Therefore, if the same department, the question for inquiry is not changed and the contents of the inquiry are limited.

  Therefore, an object of the present invention is to provide an inquiry support technology capable of outputting appropriate inquiry items from a large number of inquiry items.

  In order to solve the above-described problems, an inquiry support system according to an aspect of the present invention includes a reception unit that receives user input, an inquiry holding unit that holds a plurality of inquiry items, a first function, and a second function. Output based on at least one of the functions, an output determining unit that determines whether or not to output an inquiry item, a sending unit that outputs an inquiry item based on the determination of the output determining unit, and an output based on the first function A first evaluation input unit for receiving first evaluation information for the inquiry item; and a second evaluation input unit for receiving second evaluation information for the inquiry item output based on the second function. The output determination unit adjusts a value for determining the presence or absence of the output of the inquiry items or the priority of the output based on the evaluation information of at least one of the first evaluation information and the second evaluation information.

  Another aspect of the present invention is also an inquiry support system. This inquiry support system includes an acceptance unit that accepts user input, an inquiry holding unit that holds a plurality of inquiry items, an output determination unit that determines whether or not to output an inquiry item, and an inquiry based on the determination of the output determination unit A sending unit that outputs items, and a display unit that displays information received from the sending unit. The interview holding unit holds the inquiry items in association with the state names, the reception unit receives the answers of the examinees for the output interview items, and the display unit does not output the interview items associated with the state names The non-output display area to indicate is displayed, and the accepting unit accepts an input for outputting an inquiry item that has not been output via the non-output display area.

  Yet another embodiment of the present invention is also an inquiry support system. This inquiry support system includes a reception unit that accepts user input, an inquiry holding unit that holds a plurality of interview items and their options, an output determination unit that determines whether or not to output an inquiry item, and a determination of an output determination unit A sending unit that outputs an inquiry item based on the information and a display unit that displays information received from the sending unit. The interview holding unit holds a plurality of interview items and their options in association with a flowchart for deriving the state of the examinee, and the reception unit receives the answer of the examinee for the output interview items from the options, The display unit displays a flowchart according to the answer of the examinee, and the reception unit receives an input for outputting an inquiry item that is not output and is associated with the flowchart.

  Yet another embodiment of the present invention is also an inquiry support system. The interview support system includes a reception unit that receives user input, an interview holding unit that holds a plurality of interview items, a selection unit that selects a group of interview items among a plurality of interview items based on a predetermined condition, An output determination unit that determines the presence or absence of output of each interview item in the group of interview items selected by the selection unit, a sending unit that outputs the interview item determined to be output by the output determination unit, and each in the output determination unit A machine learning unit that performs machine learning in order to adjust a value for determining the priority of the output of the inquiry items. The output determination unit determines the priority of the output of the inquiry item related to the condition of the examinee, the first output determination unit that determines whether or not the inquiry item related to the serious condition of the examinee is output, and performs the inquiry based on the priority order A second output determining unit that determines whether or not the matter is output. The accepting unit accepts the evaluation information for the outputted inquiry item, and the machine learning unit causes the second output determining unit to perform machine learning based on the evaluation information for the inquiry item regarding the condition of the examinee.

  According to the present invention, it is possible to provide an inquiry support technology capable of outputting appropriate inquiry items from a large number of inquiry items.

It is a figure for demonstrating the structure of the inquiry assistance system concerning an Example. It is a figure for demonstrating the transition of the screen displayed on an output part in an inquiry assistance system, and is a figure which shows the screen which inputs a patient's basic information and symptom information at the time of the start of an inquiry. It is a figure for demonstrating the transition of the screen displayed on an output part in an inquiry assistance system, and is a figure which shows the screen which inputs the reply with respect to an inquiry matter. It is a figure which shows the example of a screen which displays the result of having output the inquiry matter regarding a serious condition. It is a figure which shows the example of a screen which displays the result of having output the inquiry matter regarding the state which appears frequently. It is a figure which shows the example of a screen for inputting the evaluation information with respect to the inquiry matter regarding the output serious condition. It is a figure which shows the example of a screen for inputting the evaluation with respect to the result which output the inquiry matter regarding the state which appears frequently. It is a figure which shows the example of a screen of the flowchart for derivation | leading-out of a predetermined disease. It is a figure which shows the flowchart of the process which determines the priority of an inquiry matter. It is a figure which shows the flowchart of the process which outputs an inquiry matter. It is a figure which shows the flowchart of the process which performs machine learning to an output determination part.

  FIG. 1 is a diagram for explaining a configuration of an inquiry support system 10 according to the embodiment. The inquiry support system 10 automatically outputs the inquiry items when diagnosing the examinee, supports the inquiry of medical staff such as doctors and nurses, and reduces the burden on the medical staff. In the inquiry support system 10, the medical staff may input the answer of the examinee with respect to the inquiry item, or the examinee may input the answer with respect to the inquiry item. In the following description, the term “user” includes a medical worker or a medical examinee.

  The inquiry support system 10 includes an input unit 20, a processing unit 22, and an output unit 24. The input unit 20 is a means for inputting a user operation, for example, an icon displayed on a touch panel display, and is a keyboard, a mouse, or a voice input system connected to the processing device.

  The output unit 24 outputs information related to the inquiry output from the processing unit 22. The output unit 24 may be a display that displays information related to an inquiry as an image, and also functions as a display unit. The input unit 20 and the output unit 24 may be configured by a single touch panel display, and are connected to the processing unit 22. In the modified example, the output unit 24 may be a printing machine that prints information related to an inquiry on paper. In addition, when outputting on paper, the output unit 24 may change the total number of questions to be output according to the number of printed sheets designated by the user. The number of printed sheets designated by the user is an index indicating the output amount of the inquiry item.

  The processing unit 22 executes an inquiry process for outputting an inquiry item, and executes a learning process for performing machine learning regarding the priority order of the inquiry items. The processing unit 22 calculates the priority order of the interview items to be output based on the user's symptom information, and determines whether or not the interview items are output based on the priority order of the interview items. In the processing unit 22, each element described as a functional block for performing various processes can be configured by a circuit block, a memory, and other LSIs in terms of hardware, and loaded into the memory in terms of software. It is realized by the program. Therefore, it is understood by those skilled in the art that these functional blocks can be realized in various forms by hardware only, software only, or a combination thereof, and is not limited to any one.

  By the way, in a website such as a search site, the priority order of items to be output is calculated based on a search word, and an amount of search results that can be displayed in one page is displayed based on the priority order. In the processing unit 22 of the embodiment, it is difficult to perform machine learning for the process of evaluating the priority of inquiry and maintain a high recall rate unlike the case of machine learning of the function of evaluating the display priority of a search site or the like. High degree. In the function that evaluates the display priority of search sites, etc., the person who enters the search term and the person who evaluates the quality of the search results are the same and can be completed by only one person. This is because the search accuracy is not so high in order to reach the level that the function satisfies the minimum requirement for the user.

  On the other hand, in order to realize the function that the processing unit 22 of the embodiment evaluates the priority order of the inquiry items, it is a function that excludes serious diseases and serious conditions, or a function that diagnoses frequent diseases, Ability to make decisions according to flow charts, a function to evaluate a semantic qualifier, a function to evaluate the necessary complications for treatment, and a function to evaluate the need for urgent treatment There are multiple functions, such as the ability to evaluate whether or not to return home, and the output accuracy of the inquiry items required for each function is different. For example, in order to exclude serious diseases and serious conditions, it is necessary to answer multiple questions, and the accuracy of outputting appropriate questions as a function to exclude serious diseases is high. Is desired. In the following explanation, the serious disease and serious condition of the examinee are collectively referred to as “severe condition”.

  Moreover, as a function for determining the priority order of the inquiry items, the place where the conditions of the two users, the examinee and the medical staff, and the function to be performed are different from the normal priority determination method. In addition, an inquiry regarding a serious condition is meaningful if an answer “No” is entered in an inquiry that suggests a serious illness, and even an inquiry that is likely to be answered “No”. It is necessary to output. On the other hand, since the second output determination unit 36 is a function for making an inquiry about a high frequency disease, it is often the case that an answer “No” is inputted to an inquiry suggesting a certain disease, It is not desirable to output an interview that is likely to be answered “No” because it means that it did not make sense. As described above, when the examinee answers “No” to the inquiry item, the evaluation of whether or not the output of the inquiry item is appropriate differs depending on the function.

  In order to perform a single function, such as excluding severe conditions, it is often necessary to answer multiple questions. In addition, in one example, the question to confirm related symptoms such as “Tell me about symptoms other than headache” is a serious problem in order to realize the function of diagnosing “sinusitis,” a common disease. One question may be conducted for the purpose of implementing a plurality of functions, as asked for the purpose of realizing the function of excluding the disease “cerebral infarction”.

  Further, the output accuracy of the minimum necessary interview items as a function for evaluating a frequently occurring disease is different from the output accuracy of the minimum necessary interview items as a function for excluding serious diseases. If a health care worker usually sees 50 patients a day, even if an inquiry regarding the exclusion of a serious disease is conducted with an accuracy of 98% (2% error), it will be repeated once a day. This means that it cannot play a role as a function of excluding serious diseases, and the medical risks of medical workers are greatly increased.

  Note that it is difficult to create a rule for outputting an inquiry item that practically excludes a severe condition with 100% accuracy. Therefore, when outputting the result of whether or not the serious condition has been excluded, the inquiry support system 10 clearly indicates the criteria for evaluating the serious condition to the health care worker, and how the examinee evaluates by what criteria. It is important to clearly indicate which items have been passed or have not been evaluated, and it is also important to pair the display method with feedback. In addition, it is very labor intensive for the inquiry support system to provide these feedbacks in a place different from the clinical workflow, and the cost of creating teacher data used for machine learning is very high. In addition, there is a limit to the number of questions that can be answered in a concentrated manner by the examinee. For example, the number is limited to about 3 for the questionnaire and about 20 for the screen display. In addition, since it is often difficult for the interview support system 10 to output all the interview items due to the limitations of paper and screen space, it is necessary to execute each function of each interview item of the past medical staff. It is important to extract and output a part of the feedback based on the feedback of the importance of the person and the answers of past examinees. At that time, it is also necessary to output so that the necessary questions for diagnosis and the evaluation of the possibility that the serious condition is hidden are performed as reliably as possible. In addition, consistency is required for each interview. For example, asking a male or elderly patient about the possibility of pregnancy gives the patient an incongruity and a strong sense of discomfort.

  According to the inventor's research results, the accuracy with which the processing unit 22 outputs an appropriate inquiry item may have been a recall of 60% to 80% before evaluation information is obtained for each function and machine learning is performed. In addition, it has been insufficient for realizing a function requiring a reproducibility close to 100%, such as evaluation of serious diseases. Therefore, the interview support system 10 outputs different interview items, outputs the interview items according to the different functions, inputs the answers to the interview items, displays the answer contents, and checks the validity of the question items based on the answer contents. It was possible to smoothly input five pieces of evaluation information, efficiently collect teacher data related to the determination of clinical status, and correctly perform machine learning.

  The processing unit 22 includes a reception unit 26, an inquiry holding unit 28, an extraction unit 30, a selection unit 32, a first output determination unit 34, a second output determination unit 36, a machine learning unit 38, an output holding unit 42, and a first evaluation input. Unit 43, second evaluation input unit 45, sending unit 44, and basic question holding unit 46.

  The accepting unit 26 accepts user input from the input unit 20. The inquiry holding unit 28 holds a plurality of inquiry items for asking questions to the examinee. Interviews include, for example, symptom quality, symptom start time, symptom location, symptom duration, symptom frequency, symptom severity, recurrence of symptoms, symptoms related symptoms, symptom-inducing factors, symptoms Includes hundreds of questions that ask questions about exacerbation factors, medical history, trauma history, treatment history, environmental history, family history, laboratory history, etc.

  The inquiry holding unit 28 holds answer options for the inquiry items. The inquiry items include those that can be associated with a plurality of state names and those that can be associated with one state name. In addition, an inquiry item is associated with a plurality of functions such as a function related to a serious condition and a function related to a high frequency disease. Answer options for the inquiry items are displayed together with the inquiry items, and the user can easily answer by selecting the options. The state name and function related to the inquiry item are held in association with the inquiry item. The health care worker diagnoses the possibility of corresponding to the state name by answering the question. The state name associated with the inquiry item is held in the inquiry holding unit 28 by distinguishing whether it is a serious state or a disease other than that. Note that when voice input is performed, the options are hidden and may not be displayed, but the result is the same as when the choices are input after evaluating the content of the choices based on the content of the voice input. You may be able to get

  The basic question holding unit 46 holds questions regarding basic information and symptom information of the examinee. The basic information of the examinee includes identification information, age, and gender. The symptom information of the examinee includes the main symptom of the examinee (main complaint), the nature of the main complaint, and other main complaints. The basic question holding unit 46 sends questions regarding the basic information and symptom information of the examinee to the sending unit 44 as essential questions at the start of the inquiry.

  The extraction unit 30 extracts the inquiry items held in the inquiry holding unit 28 based on the symptom information of the examinee. The extraction unit 30 extracts candidates for an inquiry item based on the symptom information of the examinee, for example, which symptom such as headache, dizziness, or common cold is main. For example, when the symptom information of the examinee is a headache, the extraction unit 30 extracts about 100 candidates for inquiry items related to the headache. More specifically, the extraction unit 30 extracts a candidate for an inquiry item based on the main complaint of the examinee and the nature of the main complaint.

  The selection unit 32 selects a group of inquiry items from the inquiry item candidates based on a predetermined selection condition. For example, when the symptom information of the examinee is a headache, the selection unit 32 selects an essential inquiry item related to the headache. In addition, if the feedback that does not require the interview items output from the medical staff at the time of execution of the past interview process is received, the selection unit 32 determines the interview items according to the input of feedback from the candidates for the interview items. except for. Further, the selection unit 32 removes unnecessary inquiry items from the inquiry item candidates based on the symptom information of the examinee and the basic information of the examinee such as the age and sex of the examinee. For example, questions that are unnecessary for men and questions that are unnecessary for women are excluded according to gender, and unnecessary questions are excluded according to age. As described above, by removing the question items that are highly uncomfortable for the user by using a plurality of selection conditions from the candidates for the question items, it is possible to prevent the medical staff from feeling uncomfortable with the output question items.

  The first output determination unit 34 and the second output determination unit 36 determine the priorities of the respective inquiry items in the group of inquiry items selected by the selection unit 32, which are learned by the machine learning unit 38 based on past feedback. It is determined by a function for machine learning such as a function, a regression function, or a similarity evaluation function, and whether or not each question item is output is determined based on certain criteria such as priority, number of pages, total number of questions, and decision boundary. In the machine learning unit 38, information about the examinee such as the sex of the examinee, age, the nature of the main complaint and the main complaint, answers to the interview items, and past feedback information of the medical staff are used as the machine learning teacher data. . The first output determination unit 34 and the second output determination unit 36 may have a function for calculating the priority order of each inquiry item when the basic information and symptom information of the examinee are input. The calculation means optimized by the machine learning unit 38 updates the calculation method of the priority order of each inquiry item in the first output determination unit 34 and the second output determination unit 36.

  In the case of executing past inquiry processing, the selection unit 32 receives the reason for making the output inquiry items unnecessary from the inquiry evaluator (in many cases, medical personnel) and the input of evaluation information. It is possible to set so as not to enter a candidate for an inquiry item under a specific selection condition in. As this selection condition, unnecessary interrogation items are excluded from the interrogation item candidates based on the basic information of the examinee such as the age and sex of the examinee. For example, it is not necessary for men to ask if they are pregnant. Such unnecessary questions and unnecessary questions for women are excluded according to gender, and unnecessary questions are excluded according to age. . In this way, by removing the less important interview items using multiple selection conditions from the candidate for the interview items, it is possible to prevent the interviewer from having a sense of incongruity from the next time. .

  The 1st output determination part 34 determines the 1st priority of the inquiry matter regarding a checkee's serious condition. The inquiry item is associated with information indicating whether the item is for diagnosing a severe condition. The first output determination unit 34 calculates the first priority order of the inquiry items based on the first function for determining the serious state of the examinee.

  Questions related to the serious condition include those necessary for inquiring serious diseases, questions necessary to evaluate whether it is possible to allow patients who may be in serious condition to go home, and progress rapidly. To evaluate necessary information to determine future treatment strategies, such as necessary interviews to evaluate the necessity of treatment to improve the condition, the presence or absence of drug allergies, renal dysfunction, and complications. The interview is included. Questions about serious conditions include the worst degree of pain, whether the pain has worsened, sudden onset, or a history of cancer or immunodeficiency. Further, the inquiry items regarding the serious condition include an inquiry item for diagnosing the serious condition of “red flag sign” when the chief complaint is headache.

  The first output determination unit 34 determines the first priority order only for the inquiry items related to the serious condition included in the group of inquiry items based on the symptom information of the examinee. The 2nd output determination part 36 determines the 2nd priority of the inquiry matter regarding not only a serious state but a checkee's state. The second output determination unit 36 diagnoses a disease group according to the symptom information of the examinee. In addition, the second output determination unit 36 determines the second priority order of the inquiry items related to the state of the examinee based on the answer to the output inquiry items during the output of the inquiry items. The second output determination unit 36 calculates the second priority based on the second function for determining the state of the examinee.

  The first output determination unit 34 determines the output of an inquiry item for diagnosing the presence or absence of the patient's serious condition. The second output determination unit 36 determines the output of an inquiry item for diagnosing a frequently occurring disease according to the symptom information and basic information of the examinee.

  Here, in the medical field, it is unlikely that the examinee is in a serious state, and the machine learning unit 38 actually inquired the examinee unintentionally and let the output decision unit machine learn with teacher data that was useful for diagnosis. In this case, the output determination unit may output an inquiry item regarding a frequently occurring state, and may make it difficult to output an inquiry item for diagnosing the presence or absence of a severe state. In addition, the health care professional must diagnose whether the patient is in a serious condition, and if he / she does not decide whether to perform a test including hospitalization or an invasive test, the patient may result in death, etc. On the other hand, diseases such as mild colds are not appropriately interviewed, and even if the diagnosis is delayed, it is unlikely that an outcome such as the death of the patient will be expected, and the possibility of a medical lawsuit is low . Therefore, the first output determining unit 34 determines the output of the inquiry item regarding the severe condition, so that the presence or absence of the severe condition can be more reliably diagnosed. In this way, by providing the first output determining unit 34 for determining the output of the inquiry item related to the severe condition separately from the second output determining unit 36 that determines the output of the inquiry item that is not limited to the severe condition, the severe condition It is possible to avoid the fact that the questioning item related to the question is buried in the questioning matter related to the frequently appearing state and is not output. In addition to the above, the output determination unit evaluates a function for excluding a severe condition, a function for diagnosing a disease, a function for executing a determination according to a flowchart, and a symptom (Semantic Qualifier) having a specific property. Determine the output of the interview items to fulfill any of the functions, such as the function, the function to evaluate the necessary complications for treatment, the function to evaluate the necessity of emergency treatment, the function to evaluate whether or not to return home Also good.

  The sending unit 44 outputs an inquiry item to the output unit 24 based on the first priority and the second priority, and causes the output unit 24 to display the inquiry item. In principle, the sending unit 44 outputs the options to the output unit 24 together with the inquiry items. The sending unit 44 outputs a predetermined number of inquiry items in descending order of the first priority and the second priority. The sending unit 44 outputs an inquiry item based on the first priority, and then outputs the inquiry item based on the second priority. By outputting the inquiry items output by the first output determination unit 34 and the inquiry items output by the second output determination unit 36 separately, the medical staff frequently asks if the inquiry item is related to a serious condition. It becomes easy to understand whether it is a question regarding the condition.

  When the inquiry item output based on the first priority and the inquiry item output based on the second priority are duplicated, the sending unit 44 outputs only one of the duplicated inquiry items. In the embodiment, since the inquiry items based on the first priority are output prior to the inquiry items based on the second priority, the inquiry items based on the second priority are deleted when the inquiry items overlap. The Thereby, it can avoid that the same inquiry matter is output. The following is an example, but an example in which one inquiry item is output by a plurality of functions will be described. For example, the question to confirm the presence or absence of a related symptom, “Do you have any of the following symptoms at the time of a headache?” Is a relatively frequent disease, even when diagnosing a serious disease. It is an important question to put on the diagnosis. For example, if a health care worker gets a high answer to the question “The limbs will not move or the force will not be applied”, it is presumed that a sudden headache is a serious disease such as a stroke. On the other hand, when a medical worker obtains an answer that the symptoms of “nose clogging” or “nasal discharge” are observed, it is estimated that the disease is a non-serious disease with a high frequency of sinusitis. As shown in these examples, the evaluation of whether an inquiry item was appropriate for realizing any function does not necessarily match the final validity of the inquiry item itself. Therefore, the learning speed can be dramatically increased by obtaining the validity evaluation for each function, and the accuracy can be dramatically improved.

  The output unit 24 displays the inquiry items output from the sending unit 44 and the options thereof, and the input unit 20 receives an answer to the inquiry item from the user. When the user selects voice input, it is assumed that an option that is assumed to be closest to the voice signal is input.

  The output holding unit 42 holds the inquiry item and the answer output by the first output determination unit 34 and the inquiry item and the answer output by the second output determination unit 36 in association with the patient identification information. . When the output of the inquiry item is completed, the medical staff can check the inquiry item and the list of answers held in the output holding unit 42 to perform diagnosis or clinical evaluation. The output unit 24 can display a list of the interview items that have been output, the answers that have been input, and the question items that have been selected but not output by the selection unit 32 after the output of the interview items and the input of the responses have been completed.

  When the medical staff confirms the list of interview items held in the output holding unit 42, if there are any interview items that are unnecessary or uncomfortable for the medical staff, the evaluation information for the output interview items and the output are output. You can enter evaluation information for questions that have not been answered. The first evaluation input unit 43 sends image information, for example, an icon for evaluation input, to which the first evaluation information for the inquiry item output by the first output determination unit 34 is input, to the transmission unit 44. The second evaluation input unit 45 sends image information, for example, an icon for evaluation input, to which the second evaluation information for the inquiry item output by the second output determination unit 36 is input, to the sending unit 44. In addition, the first evaluation input unit 43 and the second evaluation input unit 45 may send image information for inputting evaluation information for an inquiry item that has not been output to the sending unit 44.

  The input evaluation information includes, for example, negative evaluation (unnecessary), positive evaluation (necessary), and evaluation (prohibition) that prohibits output. The output unit 24 displays an icon for inputting an evaluation for each inquiry item together with the list of the output inquiry items. If a medical worker inputs an icon for inputting an evaluation, evaluation information corresponding to the operation is input. The negative or positive evaluation information and the evaluation information for prohibiting the output are used for machine learning by the machine learning unit 38 and for updating the selection condition of the selection unit 32. The selection unit 32 incorporates a rule-based expression or the like so as not to select an inquiry item corresponding to the evaluation information when the evaluation information prohibiting the output is received.

  The machine learning unit 38 causes the first output determination unit 34 and the second output determination unit 36 to perform machine learning based on the output evaluation information for the inquiry item. The machine learning unit 38 uses a learning model such as a random forest, AODE (Averaged One-Dependence Estimators), naive bayes, neural network, tree classification, deep learning, etc., evaluation information for the question and answer to the question, etc. The first output deciding unit 34 and the second output deciding unit 36 are updated using the information regarding the examinee as teacher data. When the first output determination unit 34 is a function for making an inquiry about a serious condition, it is desirable that the first output determination unit 34 outputs an appropriate inquiry item with 100% accuracy. Select a white box type classifier (a classifier that can track the process until the classification prediction result is calculated and output for the input) as a learning model with a smaller number of features, Cross validation may be performed with a plurality of learning models and various feature amounts, and the accuracy may be compared and displayed, and an appropriate model may be selected. When the accuracy of machine learning is insufficient, a plurality of machine learning models are selected and various feature quantities are selected, and the accuracy, recall, relevance rate, F value, ROC value, etc. (these The accuracy may be increased by automatically selecting machine learning based on the value (hereinafter simply referred to as “accuracy”) or by introducing a rule base. Note that only one of the first output determination unit 34 and the second output determination unit 36 may be updated by the machine learning unit 38 using machine learning.

  The machine learning unit 38 is based on the evaluation information of at least one of the first evaluation information for the inquiry item output by the first output determination unit 34 and the second evaluation information for the inquiry item output by the second output determination unit 36. The parameter value of the function for machine learning for determining whether or not to output the inquiry item is adjusted, or the parameter value of the function for machine learning for determining the priority of the inquiry item is adjusted.

  For example, the machine learning unit 38 includes two machine learning units (a first machine learning unit and a second machine learning unit), and the evaluation information for the inquiry items output by the first output determination unit 34 is the first machine. It is sent to the learning unit and used for machine learning of the first output determination unit 34. The first machine learning unit performs machine learning based on the first evaluation information as to whether the inquiry items output by the function of excluding a severe condition are appropriate, and updates the first output determination unit 34. In addition, the second machine learning unit performs machine learning based on the second evaluation information indicating whether the inquiry item is appropriate in the “function for diagnosing a certain disease group”, and updates the second output determination unit 36. That is, the second evaluation information for the inquiry item output by the second output determination unit 36 is sent to the machine learning unit 38 and used for machine learning of the second output determination unit 36. In this manner, the first output determination unit 34 and the second output determination unit 36 can be divided to perform machine learning for each function.

  For the machine learning of the second output determination unit 36, many types of input feature values are set, and among the questions related to headaches, the doctors actually need the doctor who is ranked in the top 20 in the second priority. It was converged to a state where 66% of the interview items ranked in the top 20 were judged to be. As the teacher data of the second output determination unit 36, the input feature amount includes the age of the examinee, the nature of the headache, the duration, the headache site, related symptoms, and past feedback information.

  About the machine learning of the 1st output determination part 34, the kind of input feature-value was set less than the 2nd output determination part 36, and it corresponded with the high-order inquiry matter which the doctor judged to be necessary with 100% accuracy. As the teacher data of the first output determination unit 34, the input feature amount includes the age of the examinee (whether it is 50 years or older), accompanying symptoms (presence of fever), and past feedback information.

  FIG. 2 is a diagram for explaining the transition of the screen displayed on the output unit 24 in the inquiry support system 10 and shows a screen for inputting the basic information and symptom information of the examinee at the start of the inquiry. FIG. 2 shows an example of a screen used as a medical department at an outpatient department or the like, and shows an example of treating a headache examinee.

  FIG. 2A shows a screen at the start of an inquiry, and the output unit 24 requests input of the identification information of the examinee and the main symptoms (main complaint) of the examinee. When a user such as a medical worker or a patient completes the input of the patient identification information and the main symptoms of the patient, the screen transitions to the next screen shown in FIG.

  FIG. 2B shows a screen for inputting basic information of the examinee, and the output unit 24 requests input of basic information of the examinee such as age and sex, and symptoms other than the main symptoms. When the user completes input of the basic information and other symptoms of the examinee, the screen transitions to the next screen shown in FIG.

  FIG. 2 (c) shows a screen for inputting the nature of the main symptoms of the examinee entered on the screen shown in FIG. 2 (a), and the output unit 24 requests input of the properties of the main symptoms. Thereby, the nature of the main complaint of the examinee is acquired and can be reflected in the calculation of the priority order of the inquiry items by the first output determination unit 34 and the second output determination unit 36. When the user completes the input of the nature of the main symptoms of the examinee, the screen transitions to the next screen and an inquiry item is output. The symptom information of the examinee includes the main symptom of the examinee, the nature of the symptom, and other symptoms.

  FIG. 3 is a diagram for explaining the transition of the screen displayed on the output unit 24 in the inquiry support system 10 and shows a screen for inputting an answer to the inquiry item.

  The extraction unit 30 extracts candidates for an inquiry item related to headache based on the basic information and symptom information of the examinee. Next, the selection unit 32 selects a group of inquiry items from the inquiry item candidates related to headache based on a predetermined selection condition. The first output determination unit 34 calculates the first priority of each inquiry item related to the severe condition among the group of inquiry items, and determines whether or not the inquiry item is output based on the first priority.

  As shown in FIG. 3 (a), the inquiry items regarding the severe condition output based on the first priority and the options thereof are displayed. If the user inputs one of the options, the next inquiry item is displayed. By displaying the options, the user can easily answer. In addition, when the user selects a printing method, the output unit 24 determines the inquiry items within the range of one page based on the priority order determined by the first output determination unit 34 or the second output determination unit 36. To print. After it has been entered by the examinee and the answer has been entered, the first output decision unit 34 or the second output decision unit 36 determines the priority of each question based on the answer. The output unit 24 determines and prints out the inquiry items that fit within the next page. In order to fit on one page, the number of printed sheets is not only used as an index of the output amount of the inquiry item, but the output amount of the inquiry item can be set by other numerical values.

  As shown in FIG. 3 (b), the inquiry items related to frequent diseases output based on the second priority and the options are displayed. If the user inputs one of the options, the next inquiry item is displayed. At this time, the inquiry items having the highest priority are displayed, but the display order of the inquiry items is not related to the priority. When the inquiry item based on the second priority is overlapped with the inquiry item based on the first priority, the sending unit 44 skips the overlapping inquiry item and outputs the next second priority inquiry item.

  After outputting the question items having the second priority from the first to the tenth place, the first output determination unit 34 and the second output determination unit 36 determine the remaining group of the inquiry items based on the output of the output question items. To derive the priority or the degree of output required. The sending unit 44 omits the existing inquiry items, and the sum of the number of inquiry items determined to be output by the first output determination unit 34 and the number of inquiry items determined to be output by the second output determination unit 36 is 20. The questionnaire items are output based on the priority order. In this way, the inquiry items are not all output by one calculation by the output determination unit, but are output based on the re-determination of the output determination unit that is calculated by reflecting the answer of the output inquiry item. . The first output determination unit 34 or the second output determination unit 36 executes the determination of the output of the inquiry item in a plurality of stages, so that the contents of the answer to the inquiry item output in the middle can be reflected in the priority order. .

  When the output of the inquiry item is completed and all the answers are received, the output of the inquiry item is completed. The medical staff looks at the list of questions and answers and diagnoses the patient. At that time, the inquiry items and the answers will be displayed organized by function.

  FIG. 4 shows an example of a screen that displays a result of outputting an inquiry item regarding a severe condition. As shown in FIG. 4, the output unit 24 displays information on the name of the serious condition of the examinee derived by the answer. Specifically, as a result of the answer to the inquiry item, an inquiry item and an unanswered inquiry item are displayed. Although FIG. 4 shows a display example related to one state name, the output unit 24 may actually display a list of states related to the output inquiry items. As shown in FIG. 4, the output unit 24 displays information related to the interview items that are output based on the first priority, that is, the output interview items and their answers.

  A conforming icon 52, a nonconforming icon 54, and an unquestioned icon 56 are displayed at a position below the state name 50 indicating “red flag sign”. The conformity icon 52 and the nonconformity icon 54 indicate answers to the question items, and the unquestioned icon 56 indicates the number of question items that are not output. The conforming icon 52 is a conforming display area indicating that the answer conforms to the state name, the nonconforming icon 54 is a nonconforming display area indicating that the answer does not conform to the state name, and the unquestioned icon 56 is the state name. This is a non-output display area indicating that an inquiry item associated with is not output.

  In the example shown in FIG. 4, there are a total of six questions to be associated with the state name 50 indicating “red flag sign”. The number indicated by the unquestioned icon 56 is 1, which indicates that five items of inquiry related to the state name 50 indicating “red flag sign” are output and one item remains without being output. The medical staff can evaluate the number of the matching icons 52 and the non-matching icons 54 at a glance, and evaluate whether or not the interview is sufficient to correspond to the state name 50 indicating “red flag sign”. If at least one “red flag sign” is answered as “yes”, the health care professional determines that the finding suggests a serious disease. In this case, an inquiry item related to one “red flag sign” has not been entered, and the medical staff cannot finally determine whether the “red flag sign” is positive or negative.

  As shown in FIG. 4, the output unit 24 clearly indicates how much the inquiry items related to the serious condition have been output and what remains as an inquiry that has not been output yet, and additional inquiry items as necessary. The possibility of being involved in a medical accident by a health care worker can be reduced by obtaining the answer.

  When the medical staff performs an operation input to contact the unquestioned icon 56, the setting is made such that the inquiry items that are not output are displayed to the examinee. When the medical staff thinks that it is correct that the interview items that have not been output are correctly output, the interview items are additionally output from the output unit 24 by performing an operation of adding the interview items. Answers of additional questions can be obtained from the patient. In the output of this additional inquiry item, when the presence or absence of the “red flag sign” is determined, the inquiry item that has not been output is recorded as feedback information in the machine learning unit 38 so that it can be easily output from the next time. It is used for learning by the determination unit 34.

  Further, when the output unit 24 is in the middle of outputting a plurality of interview items, for example, the timing before the output of the interview items by the first output determining unit 34 is finished and the interview items are output by the second output determining unit 36 Thus, when it is considered correct to output an inquiry item that has not been output by the medical staff, the user can input to add an inquiry item so as to output the inquiry item that has not been output. Thereby, the freedom degree of an inquiry can be raised. In addition, “sudden onset”, “progressive tendency”, and the like displayed in the inquiry items are also displayed in an area that can be operated with the touch panel, and if the medical staff inputs these operations, the inquiry items can be further added.

  FIG. 5 shows an example of a screen that displays a result of outputting an inquiry item related to a frequently occurring state. Further, FIG. 5 shows a result of the second output determination unit 36 outputting an inquiry item for diagnosing a disease group related to headache.

  As shown in FIG. 5, the output unit 24 displays information related to the state name of the examinee that is derived from the answer. Specifically, the output unit 24 examines the examinee according to the state name, the result of the answer to the question, the question, and the disease. The body information that appears on the screen is displayed. The output unit 24 displays the information regarding the inquiry items output according to the second priority, including the answers to the inquiry items. Further, when outputting, when the answer to the question is essential for the diagnosis of the condition of the examinee, or when the diagnosis of the condition of the examinee is confirmed, the output unit 24 may display the fact.

  Similar to the screen example shown in FIG. 4, a conforming icon 52, a nonconforming icon 54, and an unquestioned icon 56 are displayed at a position below the state name 50 indicating muscle tension headache. In the example shown in FIG. 5, there are a total of four questions to be associated with the state name 50 indicating muscle tension headache. The number indicated by the unquestioned icon 56 is 2, which indicates that two questions that are associated with the state name 50 indicating muscle tension headache are output and two remain without being output. Since the number indicated by the conforming icon 52 is zero and the number indicated by the nonconforming icon 54 is 2, it indicates that the examinee is unlikely to have a muscle tension headache.

  In this way, the health care professional can diagnose whether the possibility of corresponding to the state name is high by looking at the number of the conforming icons 52 and the nonconforming icons 54. Further, if the medical worker inputs the unquestioned icon 56, it is possible to add and output the remaining inquiry items without being output, and prompt the examinee to additionally input. In addition, when a medical staff adds an inquiry item, feedback is given to the machine learning unit 38, and the second output determining unit 36 can be used as teacher data for machine learning.

  FIG. 6 shows an example of a screen for inputting evaluation information for an inquiry item related to a severe condition. As shown in FIG. 6, the output unit 24 displays an inquiry item, options and answers to the inquiry item, presence / absence of output of the inquiry item, and means for inputting the evaluation.

  The output presence / absence icon 58 indicating the presence / absence of output indicates that the output is performed when the check symbol is displayed, and indicates that the output is not performed when the check symbol is not displayed. In FIG. 6, the question items “sudden onset” and “worsening headache” are output, and the question item “worst headache” is not output.

  The first evaluation input unit 60a is an icon that allows a feedback to be input when a medical worker has a sense of incongruity in the output inquiry item, for example, when the medical worker has a sense of incongruity in the output order. The second evaluation input unit 60b is an icon that allows a feedback to be input when the medical staff determines that the outputted inquiry item is unnecessary.

  The 3rd evaluation input part 60c is an icon which can input a feedback, when a medical worker judges that the inquiry matter which is not output is required. In the evaluation information input by the third evaluation input unit 60c, in FIG. 6, positive evaluation information is input for the inquiry item of “worst headache” that has not been output this time, and the machine learning unit 38 Sent to be used for machine learning by the first output determination unit 34. The 1st output determination part 34 is adjusted so that the priority of the inquiry matter with respect to input information may go up by performing machine learning with the positive evaluation information input by the 1st evaluation input part 60a. In addition, the achievement rate by the cross validation of the teacher data is updated. When the accuracy of cross-validation of teacher data related to a severe condition is less than 99%, the accuracy is evaluated by selecting multiple machine learning models and various feature quantities, and machine learning is automatically selected based on the accuracy. Therefore, it is possible to increase the accuracy.

  In this way, the machine learning unit 38 can cause the first output determination unit 34 to perform machine learning with the evaluation information for the inquiry item related to the serious condition, thereby outputting the inquiry item related to the appropriate serious condition. In addition, by dividing the feedback path from the inquiry items regarding the frequently appearing state, the evaluation of the inquiry items regarding the severe condition can be appropriately reflected.

  FIG. 7 shows an example of a screen for inputting an evaluation on the result of outputting the inquiry items regarding the frequently appearing state. As shown in FIG. 7, the output unit 24 displays an inquiry item, an answer to the inquiry item, whether the inquiry item is output, and a means for inputting an evaluation.

  The output presence / absence icon 58 indicating the presence / absence of output indicates that the output is performed when the check symbol is displayed, and indicates that the output is not performed when the check symbol is not displayed.

  The first evaluation input unit 60a is an icon that allows a feedback to be input when a medical worker has a sense of incongruity in the output inquiry item, for example, when the medical worker has a sense of incongruity in the output order. The evaluation information input by the first evaluation input unit 60a, negative evaluation information for the “duration” question item in FIG. 7 is input, and is sent to the machine learning unit 38 to determine the second output. Used for machine learning of the unit 36. The second output determination unit 36 performs machine learning with the negative evaluation information input by the first evaluation input unit 60a, so that the priority of the inquiry item with respect to the input information is lowered at the next execution of the machine learning. Adjusted to

  The second evaluation input unit 60b is an icon that allows a feedback to be input when the medical staff determines that the outputted inquiry item is unnecessary. The evaluation information input by the second evaluation input unit 60b is sent to the selection unit 32, and a predetermined selection condition is changed.

  The 3rd evaluation input part 60c is an icon which can input a feedback, when a medical worker judges that the inquiry matter which is not output is required. The evaluation information input by the third evaluation input unit 60c, in FIG. 7, positive evaluation information is input for the inquiry item “influence on daily life”, is sent to the machine learning unit 38, and the first This is used for machine learning of the two-output determination unit 36. The second output determination unit 36 performs machine learning with the positive evaluation information input by the first evaluation input unit 60a, so that the priority of the inquiry items with respect to the input information is increased during the next machine learning execution. To be adjusted.

  In this way, the machine learning unit 38 can appropriately output the inquiry item related to the disease by causing the second output determination unit 36 to perform machine learning with the evaluation information for the inquiry item related to the disease. In addition, the inputted answer result of the examinee is displayed along the following three clinical processes. Specifically, firstly, a process for interrogating whether an emergency disease or serious condition is correctly excluded, second, a process for conducting a necessary interrogation in accordance with a clinical algorithm, and third, a result of the interrogation A process is performed in which the correct diagnosis is performed by comparing the characteristics of each disease.

  FIG. 8 shows a screen example of a flowchart for deriving a predetermined state. The output unit 24 displays the outputted inquiry items, options, and answers held in the output holding unit 42 as a flowchart. In order to derive the state of the diagnostician, the flowchart displays the inquiry items, their options, and the presence or absence of a predetermined state according to the options, and is held in the inquiry holding unit 28. In the flowchart held in the inquiry holding unit 28, the outputted inquiry items held in the output holding unit 42 and the answers are displayed together.

  FIG. 8 shows a flowchart for deriving the presence or absence of migraine, and the derivation result that there is migraine is displayed by dividing the case according to the answer. A medical worker can determine the presence or absence of migraine by looking at the flowchart regarding migraine. The flowchart is displayed for various state names, not limited to migraine.

  As shown in FIG. 8, the question “Q3. Is daily activity hindered by headache?” Is not output. The medical staff can perform an input to output an inquiry item that has not been output by performing an operation of contacting the inquiry item that has not been output and is displayed in the flowchart. Also, by pressing the add icon 59, it is possible to add an inquiry so that an inquiry can be executed from this flowchart until reaching some conclusion. Thereby, the freedom degree of an inquiry can be raised. Further, when a medical staff adds an inquiry item, feedback is given to the machine learning unit 38, which can be used as teacher data for causing the second output determining unit 36 to perform machine learning.

  FIG. 9 shows a flowchart of processing for determining the priority order of the inquiry items. A basic question for inputting the symptom information and basic information of the examinee held by the basic question holding unit 46 is displayed from the output unit 24, and the user inputs the symptom information and basic information of the examinee (S10).

  The extraction unit 30 extracts candidate interview items based on the symptom information of the examinee (S12). Thereby, the inquiry items regarding the main symptoms of the examinee are extracted. The selection unit 32 selects a group of inquiry items from the inquiry item candidates based on a predetermined selection condition (S14). Thereby, unnecessary interrogation items can be removed based on the basic information of the examinee.

  The first output determination unit 34 calculates the first priority of the inquiry items related to the serious condition from the group of inquiry items, determines the presence / absence of output of the inquiry items based on the first priority, and the second output determination unit 36 Calculates the second priority of the inquiry items related to the disease from the group of inquiry items, and determines whether or not to output the inquiry items based on the first priority (S16). Thereby, the inquiry item regarding a serious condition and the inquiry item regarding a frequent state or a frequent disease can be output separately.

  FIG. 10 shows a flowchart of a process for outputting an inquiry item. The output unit 24 outputs an inquiry item based on the first priority (S20). The output unit 24 determines whether or not all the inquiry items whose first priority is from 1st to 10th are output up to a predetermined first order (S22), and if not output (N in S22), The output of the inquiry items based on 1 priority is repeated (S20).

  If an inquiry item based on the first priority is output up to a predetermined first order (Y in S22), the output unit 24 outputs an inquiry item based on the second priority (S24). It should be noted that the inquiry item that has already been output based on the first priority is skipped, and the output unit 24 outputs the next inquiry item of the second priority. Further, if the sending unit 44 determines that it is not appropriate to make an inquiry based on the already obtained input result, such as a male having a history of uterine diseases, the inquiry is skipped.

  The output unit 24 determines whether or not all the inquiry items whose second priority is from 1st to 10th are output up to a predetermined second order (S26), and if not output (N of S26), The output of the inquiry items based on the two priorities is repeated (S24). If the inquiry items whose second priority order is a predetermined second order are output (Y in S26), the inquiry item output process is terminated. At that time, confirm whether other new chief complaints have been entered in the answers to the interview items according to the chief complaint, and if the new chief complaint has an interview item regarding the serious condition, Interviews may be added. By repeating these processes, it becomes possible to finish the evaluation of the questions regarding the serious condition for all symptoms of the examinee.

  FIG. 11 shows a flowchart of a process for performing machine learning in the output determination unit. The output unit 24 receives the information on the output of the interview items held in the output holding unit 42, displays the interview items related to the severe condition output according to the determination of the first output determination unit 34 (S30), and the second The inquiry items related to the disease output according to the determination of the output determination unit 36 are displayed (S32).

  The evaluator of the inquiry sees the output result of the inquiry item, inputs the evaluation information for the inquiry item regarding the serious condition, and inputs the evaluation information for the inquiry item regarding the disease (S34).

  The machine learning unit 38 performs machine learning on the first output determining unit 34 based on the evaluation information for the inquiry items regarding the severe condition (S36). The machine learning unit 38 performs machine learning on the second output determination unit 36 based on the evaluation information for the inquiry items related to the severe condition (S38). Thereby, the value for calculating the priority of the inquiry item regarding the serious condition and the inquiry item regarding the frequently appearing state can be adjusted.

  The present invention is not limited to the above-described embodiment, and various modifications such as design changes can be added based on the knowledge of those skilled in the art. The configuration shown in each figure is for explaining an example, and any configuration that can achieve the same function can be changed as appropriate, and the same effect can be obtained.

  In the embodiment, the mode in which the input unit 20 and the output unit 24 are connected to the processing unit 22 has been described, but the present invention is not limited to this mode. For example, the input unit 20, the processing unit 22, and the output unit 24 may be provided integrally with the mobile terminal. The input unit 20 and the output unit 24 may be a mobile terminal having a touch panel function, and the processing unit 22 may be a server connected to the mobile terminal by wire or wirelessly.

  In the embodiment, a mode in which the machine learning unit 38 is included in the processing unit 22 is shown, but the present invention is not limited to this mode. For example, the machine learning unit 38 may be provided in a processing device different from the processing unit 22. The processing unit 22 that outputs the inquiry items accumulates the evaluation information, and the machine learning unit 38 provided in another processing apparatus receives the accumulated evaluation information and sends it to the first output determination unit 34 and the second output determination unit 36. Make machine learning.

  In the embodiment, the first output determination unit 34 and the second output determination unit 36 calculate the priority order of the inquiry items based on the symptom information of the examinee, and determine whether or not to output the inquiry items based on the priority order. However, the present invention is not limited to this mode. For example, the first output determination unit 34 and the second output determination unit 36 may determine whether or not to output the inquiry item based on the symptom information of the examinee without calculating the priority order of the inquiry item. In addition, the first output determination unit 34 determines whether or not to output the inquiry item based on the symptom information of the examinee without calculating the priority order of the inquiry item, and the second output determination unit 36 A priority order may be calculated, and whether or not to output an inquiry item may be determined based on the priority order.

  DESCRIPTION OF SYMBOLS 10 inquiry support system, 20 input part, 22 processing part, 24 output part, 26 reception part, 28 inquiry inquiry holding part, 30 extraction part, 32 selection part, 34 1st output determination part, 36 2nd output determination part, 38 machine Learning unit, 42 output holding unit, 43 first evaluation input unit, 44 sending unit, 45 second evaluation input unit, 46 basic question holding unit, 50 state name.

  The present invention relates to an inquiry support system that outputs an inquiry item for diagnosing a patient.

Claims (7)

A reception unit that accepts user input;
An interview holding section for holding multiple interview items;
An output determining unit that determines whether or not to output the inquiry item based on at least one of the first function and the second function;
A sending unit that outputs the inquiry items based on the determination of the output determining unit;
A first evaluation input unit for receiving first evaluation information for the inquiry items output by the output determination unit based on the first function;
A second evaluation input unit for receiving second evaluation information for the inquiry item output by the output determination unit based on the second function,
The output determination unit adjusts a value for determining the presence or absence of output of the inquiry items or an output priority order based on the evaluation information of at least one of the first evaluation information and the second evaluation information. A characteristic inquiry support system. The output determining unit
A first output determination unit that determines whether or not to output the inquiry items based on the first function for determining a serious condition of the examinee;
2. An inquiry support system according to claim 1, further comprising: a second output determination unit that determines whether or not to output the inquiry item based on the second function for determining the state of the examinee. . A reception unit that accepts user input;
An interview holding section for holding multiple interview items;
An output determination unit for determining whether or not to output the inquiry items;
A sending unit that outputs the inquiry items based on the determination of the output determining unit;
A display unit for displaying information received from the sending unit,
The inquiry holding unit holds the inquiry item in association with a state name,
The reception unit receives a response of the examinee with respect to the outputted inquiry items,
The display unit displays an unoutput display area indicating that the inquiry item associated with the state name is not output,
The said reception part receives the input which outputs the said inquiry matter which is not output via the said non-output display area, The inquiry assistance system characterized by the above-mentioned. The output determining unit
A first output determination unit for determining whether or not to output the inquiry items related to the serious condition of the examinee;
The interrogation support system according to claim 3, further comprising: a second output determination unit that determines whether or not the interrogation items regarding the state of the examinee are output. A reception unit that accepts user input;
An interview holding section for holding a plurality of interview items and their options;
An output determination unit for determining whether or not to output the inquiry items;
A sending unit that outputs the inquiry items based on the determination of the output determining unit;
A display unit for displaying information received from the sending unit,
The inquiry holding unit holds a plurality of the inquiry items and the options thereof in association with a flowchart for deriving the state of the examinee,
The reception unit receives the answer of the examinee for the output of the inquiry items from the options,
The display unit displays the flowchart according to the answer of the examinee,
The said reception part receives the input which outputs the said inquiry matter which is linked | related with the said flowchart and is not output, The inquiry assistance system characterized by the above-mentioned. A reception unit that accepts user input;
An interview holding section for holding multiple interview items;
A selection unit that selects a group of interview items among the plurality of interview items based on a predetermined condition;
An output determining unit that determines whether or not to output each question item in the group of question items selected by the selection unit;
A sending unit that outputs the inquiry items whose output is determined by the output determining unit;
A machine learning unit that performs machine learning to adjust a value for determining a priority of output of each inquiry item in the output determination unit,
The output determining unit
A first output determination unit for determining whether or not to output the inquiry items related to the serious condition of the examinee;
A second output determining unit that determines the priority of the output of the interview items related to the state of the examinee, and determines whether or not the interview items are output based on the priority;
The accepting unit accepts evaluation information for the interrogated matter that has been output,
The said machine learning part makes the said 2nd output determination part machine-learn based on the evaluation information with respect to the said inquiry matter regarding a checkee's state, The inquiry support system characterized by the above-mentioned. A printing unit that prints the interview items received from the sending unit;
The reception unit receives an output amount of the inquiry item,
The inquiry support system according to claim 1, wherein the output determination unit determines the total number of the inquiry items to be output based on the output amount.

Images