Medical information processing apparatus, medical information processing method, and program

Abstract

To prevent an error in diagnosis.SOLUTION: A medical information processing apparatus of an embodiment has an acquisition unit and a prediction unit. The acquisition unit acquires behavior history information related to the history of a doctor's behavior and circumferential situation information related to a situation on the circumference of the doctor before diagnosis. The prediction unit predicts the occurrence of the doctor's diagnostic error on the basis of pre-diagnosis situation information related to a pre-diagnosis situation including at least either one of the history of behavior or the circumferential situation and diagnostic error information related to a diagnostic error in the past and a diagnostic error occurrence time situation at the occurrence of the diagnostic error.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The embodiments disclosed in this specification and the drawings relate to a medical information processing apparatus, a medical information processing method, and a program.

Background Art

[0002] A doctor diagnoses a patient. However, if the doctor is in a poor physical condition, such as having a hangover, a cold, the surrounding noise during diagnosis, or having encountered a traffic accident just before diagnosis, the diagnosis result may not be accurate. For this reason, the diagnosis result is often confirmed for its correctness, for example, in a conference among multiple doctors. However, depending on the medical facility, the conference among doctors may be omitted, or even if a conference is held, there may be minor diagnostic errors.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The problem to be solved by the embodiments disclosed in this specification and the drawings is to suppress diagnostic errors. However, the problems to be solved by the embodiments disclosed in this specification and the drawings are not limited to the above problems. The problems corresponding to the respective effects of each configuration shown in the embodiments described later can also be regarded as other problems.

Means for Solving the Problems

[0005] The medical information processing apparatus according to the embodiment includes an acquisition unit and a prediction unit. The acquisition unit acquires action history information regarding the action history of a doctor before diagnosis and peripheral situation information regarding the situation around the doctor. The prediction unit predicts the occurrence of a diagnostic error of the doctor based on pre-diagnosis situation information regarding a pre-diagnosis situation including at least one of the action history or the peripheral situation, and diagnostic error information regarding past diagnostic errors and the situation at the time of occurrence of the diagnostic errors.

Brief Description of the Drawings

[0006]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Modes for Carrying Out the Invention

[0007] Hereinafter, a medical information processing apparatus, a medical information processing method, and a program according to an embodiment will be described with reference to the drawings.

[0008] The medical information processing apparatus according to the embodiment is an apparatus that processes information related to diagnostic errors in the diagnostic results of diagnoses made by doctors. The medical information processing apparatus predicts the occurrence of a doctor's diagnostic error based on, for example, pre-diagnosis situation information related to a pre-diagnosis situation including at least one of the doctor's action history or the doctor's surrounding situation, and diagnostic error information related to past diagnostic errors and the situation at the time of occurrence of the diagnostic error.

[0009] The diagnostic error includes, for example, an error of a magnitude expressed as a so-called near miss. The medical information processing apparatus notifies the contents thereof, for example, when the occurrence of a diagnostic error is predicted, the occurrence of the situation at the time of occurrence of the diagnostic error is inferred, or the occurrence of the diagnostic error is detected.

[0010] The medical information processing apparatus according to the embodiment is provided, for example, in a room of a medical facility and is used by a doctor (hereinafter referred to as a diagnosing doctor) who has made a diagnosis. In another room of the medical facility where the medical information processing apparatus is provided, a conference room where a plurality of doctors (hereinafter referred to as a judging doctor group) hold a conference is provided. A conference server used by the judging doctor group is provided in the conference room. The medical information processing apparatus is capable of transmitting and receiving various information to and from the conference server.

[0011] (First Embodiment) FIG. 1 is a block diagram showing an example of the configuration of a medical information processing apparatus 100 according to the first embodiment. The medical information processing apparatus 100 is connected to a living environment sensor 10 and a conference server 50 via, for example, a LAN (Local Area Network), the Internet, a cellular network, a Wi-Fi network, a WAN (Wide Area Network), or the like.

[0012] The living environment sensor 10 includes a plurality of sensors that detect living environment information including information on the doctor's action history and surrounding situations. The living environment information includes, for example, action history information regarding the doctor's action history and surrounding situation information regarding the doctor's surrounding situation. Specifically, the surrounding situation information includes various situations such as the doctor's own actions and words, the reactions and voices of people around the doctor, and the sounds emitted by vehicles and animals. The living environment sensor 10 includes, for example, a surveillance camera 12, a voice recorder 14, a wearable sensor 16, and a processing unit (not shown) that processes the information detected by these sensors.

[0013] The surveillance camera 12 is, for example, a camera installed in a medical facility where the doctor stays or on the doctor's commuting route. The surveillance camera 12 outputs the image information including the captured doctor to the processing unit. The voice recorder 14 is provided, for example, in a portable terminal device possessed by the doctor. The voice recorder 14 includes, for example, a microphone and a memory. The voice recorder 14 outputs the voice information regarding the voices around the doctor collected by the microphone to the processing unit.

[0014] The wearable sensor 16 is, for example, a sensor that the doctor wears and detects biometric information regarding the doctor's body. The wearable sensor 16 detects, for example, the doctor's body temperature, pulse, respiratory rate, blood pressure, etc. as biometric information. The wearable sensor 16 outputs the detected biometric information to the processing unit. The processing unit generates image processing information regarding the doctor's actions by performing image processing on the image information transmitted by the surveillance camera 12. The processing unit generates living environment information including the generated image processing information, the voice information output by the voice recorder 14, and the biometric information output by the wearable sensor 16 and transmits it to the medical information processing device 100.

[0015] The conference server 50 includes, for example, a communication interface, an input / output interface, an arithmetic unit, etc. The conference server 50 transmits and receives signals to and from external devices such as the medical information processing device 100 via the communication interface. The conference server 50 causes the input / output interface to display or output various information generated by the arithmetic unit as voice.

[0016] The conference server 50 transmits determination result information including the determination result of diagnostic errors input by the panel of judging doctors to the medical information processing device 100. For example, the conference server 50 transmits error information input by the panel of judging doctors who have determined that there is a diagnostic error in the diagnosis result of the diagnosing doctor to the medical information processing device 100. For example, the conference server 50 transmits error-free information input by the panel of judging doctors who have determined that there is no diagnostic error in the diagnosis result of the diagnosing doctor to the medical information processing device 100.

[0017] The medical information processing device 100 includes, for example, a communication interface 110, an input interface 120, a display 130, a processing circuit 140, and a memory 150. The communication interface 110 communicates with external devices such as the living environment sensor 10 and the conference server 50 via a network NW such as a LAN (Local Area Network). The communication interface 110 includes, for example, a communication interface such as a NIC (Network Interface Card).

[0018] The input interface 120 receives various input operations from a user such as a doctor, converts the received input operations into electrical signals, and outputs them to the processing circuit 140. For example, when an input operation is performed by the user, the input interface 120 generates information corresponding to the input operation. The input interface 120 outputs the generated information corresponding to the input operation to the processing circuit 140.

[0019] The input interface 120 includes, for example, a mouse, a keyboard, a trackball, a switch, a button, a joystick, a touch panel, etc. The input interface 120 may be a user interface that receives voice input such as a microphone. When the input interface 120 is a touch panel, the input interface 120 may also have the display function of the display 130.

[0020] Note that in this specification, the input interface is not limited to only those equipped with physical operation components such as a mouse and a keyboard. For example, an electrical signal processing circuit that receives an electrical signal corresponding to an input operation from an external input device provided separately from the apparatus and outputs this electrical signal to the control circuit is also included in the examples of the input interface.

[0021] The display 130 is a display unit that displays various types of information. For example, the display 130 displays an image generated by the processing circuit 140, a GUI (Graphical User Interface) for receiving various input operations from the user, and the like. For example, the display 130 is an LCD (Liquid Crystal Display), a CRT (Cathode Ray Tube) display, an organic EL (Electro Luminescence) display, or the like.

[0022] The processing circuit 140 includes, for example, an acquisition function 141, a generation function 142, a storage function 143, and a prediction function 144. The processing circuit 140 realizes these functions, for example, by a hardware processor (computer) executing a program stored in the memory (storage circuit) 150.

[0023] The hardware processor means, for example, a CPU, a GPU (Graphics Processing Unit), an application specific integrated circuit (ASIC), a programmable logic device (for example, a simple programmable logic device (SPLD) or a complex programmable logic device (CPLD)), a field programmable gate array (FPGA), or the like circuitry.

[0024] Instead of storing the program in the memory 150, it may be configured to directly incorporate the program into the circuitry of the hardware processor. In this case, the hardware processor realizes its functions by reading and executing the program incorporated in the circuitry. The above program may be stored in the memory 150 in advance, or may be stored in a non-transitory storage medium such as a DVD or a CD-ROM, and may be installed from the non-transitory storage medium into the memory 150 when the non-transitory storage medium is mounted on a drive device (not shown) of the medical information processing apparatus 100.

[0025] The memory 150 stores a pre-diagnosis situation database (hereinafter referred to as DB) 151 and a diagnosis error prediction model 152. The memory 150 is an example of a storage unit. The pre-diagnosis situation DB 151 is a DB that includes a plurality of pieces of living environment information. The living environment information stored in the pre-diagnosis situation DB 151 is used as a pre-diagnosis situation before a doctor makes a diagnosis. The diagnosis error prediction model 152 is a doctor's action model generated based on the pre-diagnosis situation (living environment information) when a diagnosis error has occurred in the doctor's diagnosis result. The action model will be described later.

[0026] The hardware processor is not limited to being configured as a single circuit, and may be configured as one hardware processor by combining a plurality of independent circuits so as to realize each function. Alternatively, a plurality of components may be integrated into one hardware processor so as to realize each function.

[0027] The memory 150 is realized by, for example, a semiconductor memory element such as a RAM or a flash memory, a hard disk, or an optical disk. These non-transitory storage media may be realized by other storage devices connected via a communication network such as a NAS (Network Attached Storage) or an external storage server device. The memory 150 may include non-transitory storage media such as a ROM (Read Only Memory) or a register.

[0028] The acquisition function 141 acquires, as the pre-diagnosis situation, the action history information and the surrounding situation information included in the living environment information based on the living environment information transmitted by the living environment sensor 10 and received by the communication interface 110. The acquisition function 141 also acquires the date and time when each item of the acquired pre-diagnosis situation was detected by the living environment sensor 10. The acquisition function 141 is an example of an acquisition unit.

[0029] The acquisition function 141 acquires the determination result information transmitted by the conference server 50. Based on the acquired determination result information, the acquisition function 141 acquires information on whether there is a diagnosis error in the diagnosis result of the diagnosing doctor. For example, when the acquired determination result information is error-containing information, the acquisition function 141 acquires information indicating that there is a diagnosis error, and when the acquired determination result information is error-free information, the acquisition function 141 acquires information indicating that there is no diagnosis error.

[0030] The generation function 142 generates an action model that models the actions of the doctor on the diagnosis date based on the living environment information acquired by the acquisition function 141. For example, the generation function 142 converts the doctor included in the image processing information included in the living environment information into a bone model, and generates information on the movement (action) of the doctor included in the action model. Instead of generating a bone model, the generation function 142 may generate information on the movement of the doctor included in the action model by image processing.

[0031] The storage function 143 stores and accumulates the pre-diagnosis situation acquired by the acquisition function 141 in the memory 150. For example, each time a diagnosing doctor makes a diagnosis, the storage function 143 attaches a diagnosis ID to the diagnosis, and stores and accumulates the pre-diagnosis situation (including the date and time) for each diagnosis ID to generate a pre-diagnosis situation DB 151, or updates the generated pre-diagnosis situation DB 151 (hereinafter, the generation of the DB includes the update of the DB) and stores and accumulates it in the memory 150. The storage function 143 associates the action model generated by the generation function 142 with the pre-diagnosis situation and stores it in the pre-diagnosis situation DB 151. The storage function 143 is an example of a storage unit.

[0032] When the acquisition function 141 acquires error-related information, the storage function 143 extracts the pre-diagnosis situation of the diagnostician who made the diagnosis determined to be a diagnostic error on the day the diagnosis was made. The storage function 143 stores and accumulates in the memory 150 the action model associated with the extracted pre-diagnosis situation as a diagnostic error prediction model 152.

[0033] The pre-diagnosis situation and the action model included in the diagnostic error prediction model 152 may be one set or multiple sets. The diagnostic error prediction model 152 relates to past diagnostic errors and the situation at the time of occurrence of diagnostic errors. The diagnostic errors and the situation at the time of occurrence of diagnostic errors are associated with each other and stored in the memory 150. The diagnostic error prediction model 152 is an example of diagnostic error information.

[0034] Based on the pre-diagnosis situation information regarding the pre-diagnosis situation including the living environment information before the diagnostician makes a diagnosis and the diagnostic error information regarding past diagnostic errors and the situation at the time of occurrence of diagnostic errors, the prediction function 144 predicts the diagnostic error of the diagnostician. The prediction function 144 reads and acquires, for example, the past diagnostic error information stored in the memory 150.

[0035] The prediction function 144 calculates, for example, the similarity between the action model generated by the generation function 142 and the diagnostic error prediction model 152 stored in the memory 150. When the calculated similarity is equal to or greater than a threshold value, the prediction function 144 predicts that there may be a diagnostic error in the diagnosis made by the diagnostician. The prediction function 144 is an example of a prediction unit.

[0036] Next, the processing in the medical information processing apparatus 100 will be described. Before making a diagnosis, the diagnosing doctor has the living environment information detected by the living environment sensor 10, and the living environment sensor 10 transmits the living environment information to the medical information processing apparatus 100. After making a diagnosis, the diagnosing doctor provides it to the judging doctor group. The judging doctor group conducts a conference regarding the presence or absence of a diagnostic error in the diagnosis result of the diagnosing doctor, and inputs the judgment result information to the conference server 50. The conference server 50 transmits the input judgment result information to the medical information processing apparatus 100.

[0037] After the living environment information is transmitted by the living environment sensor 10 and the judgment result information is transmitted by the conference server 50, the medical information processing apparatus 100 generates a diagnostic error prediction model 152 or predicts the possibility of a diagnostic error. Hereinafter, the process of generating the diagnostic error prediction model 152 and the process of predicting the possibility of a diagnostic error by the medical information processing apparatus 100 will be described. FIGS. 2 and 3 are flowcharts showing an example of the processing of the medical information processing apparatus 100 according to the first embodiment.

[0038] First, the process of generating the diagnostic error prediction model 152 will be described. FIG. 2 shows the flow of the process of generating the diagnostic error prediction model 152. In the medical information processing apparatus 100, first, the acquisition function 141 acquires the pre-diagnosis situation of the diagnosing doctor based on the living environment information transmitted by the living environment sensor 10 and received by the communication interface 110 (step S101).

[0039] Subsequently, the storage function 143 stores by generating a pre-diagnosis situation DB 151 based on the pre-diagnosis situation acquired by the acquisition function 141 and storing it in the memory 150 (step S103). Subsequently, the generation function 142 generates a diagnostic doctor's action model based on the pre-diagnosis situation acquired by the acquisition function 141 (step S105). Subsequently, the storage function 143 associates the pre-diagnosis situation acquired by the acquisition function 141 with the action model generated by the generation function 142 (step S107), generates a pre-diagnosis situation DB 151, and stores it in the memory 150 for storage (step S109).

[0040] Subsequently, the acquisition function 141 acquires determination result information transmitted by the conference server 50 (step S111). Subsequently, the storage function 143 determines whether there is a diagnostic error in the diagnostic result of the diagnostic doctor based on whether the determination result information acquired by the acquisition function 141 is error-containing information or non-error information (step S113).

[0041] When the acquisition function 141 acquires error-containing information and the storage function 143 determines that there is a diagnostic error in the diagnosis by the diagnostic doctor, the storage function 143 extracts the pre-diagnosis situation on the day when the diagnostic doctor who made the diagnosis determined to be a diagnostic error made that diagnosis. The storage function 143 generates a diagnostic error prediction model 152 by associating the extracted pre-diagnosis situation with the action model generated by the generation function 142, stores it in the memory 150, and stores it (step S115). Thus, the medical information processing apparatus 100 ends the process shown in FIG. 2.

[0042] On the other hand, in step S111, when the acquisition function 141 acquires non-error information and the storage function 143 determines that there is no diagnostic error in the diagnostic result of the diagnostic doctor, the medical information processing apparatus 100 simply ends the process shown in FIG. 2.

[0043] Next, a process for predicting the possibility of a diagnostic error will be described. FIG. 3 shows the flow of the process for predicting the possibility of a diagnostic error. The medical information processing apparatus 100 first acquires the pre-diagnosis situation of the diagnosing doctor based on the living environment information transmitted by the living environment sensor 10 and received by the communication interface 110 in the acquisition function 141 (step S121).

[0044] Subsequently, the generation function 142 generates an action model of the diagnosing doctor based on the pre-diagnosis situation acquired by the acquisition function 141 (step S123). The processes of step S121 and step S123 may directly utilize the processes of step S101 and step S105 in the process of generating the diagnostic error prediction model 152 shown in FIG. 2.

[0045] Subsequently, the prediction function 144 calculates the similarity (hereinafter, the comparison target action model) between the action model generated by the generation function 142 (hereinafter, the determination target action model) and the action model stored in the diagnostic error prediction model 152 (step S125). When the diagnostic error prediction model 152 includes a plurality of action models, the similarity between the determination target action model and each of the plurality of comparison target action models is calculated. The prediction function 144 calculates the similarity by, for example, pattern matching.

[0046] Subsequently, the prediction function 144 determines whether the calculated similarity is equal to or greater than a predetermined threshold (step S127). If it is determined that the calculated similarity is equal to or greater than the predetermined threshold, the prediction function 144 predicts that there is a possibility of a diagnostic error (step S129). If it is determined that the calculated similarity is less than the predetermined threshold, the prediction function 144 predicts that there is no possibility of a diagnostic error (step S131).

[0047] When a plurality of comparison target behavior models are included in the diagnostic error prediction model 152, the prediction function 144 calculates the similarity between the determination target behavior model and each of the plurality of comparison target behavior models, and determines whether any of the calculated similarities is equal to or greater than a threshold value. When any of the calculated similarities is equal to or greater than the threshold value, the prediction function 144 predicts that there may be a diagnostic error, and when all of the calculated similarities are less than the threshold value, the prediction function 144 predicts that there is no possibility of a diagnostic error.

[0048] Subsequently, the prediction function 144 transmits the prediction result of the possibility of a diagnostic error to the conference server 50 (step S133). Thus, the medical information processing apparatus 100 ends the processing shown in FIG. 3. The conference server 50 provides the transmitted prediction result to the judging doctor group. The judging doctor group determines the presence or absence of a diagnostic error in the diagnostic result by the diagnosing doctor with reference to the provided prediction result.

[0049] The prediction function 144 determines the possibility of a diagnostic error based on whether the similarity between the determination target behavior model and the comparison target behavior model is equal to or greater than a threshold value. However, the prediction function 144 may determine the possibility of a diagnostic error in other manners. Further, the prediction function 144 may determine the possibility of a diagnostic error as a percentage. In this case, for example, it may be predicted that the higher the similarity between the determination target behavior model and the comparison target behavior model, the higher the possibility (percentage) of a diagnostic error.

[0050] The medical information processing apparatus 100 according to the first embodiment generates a determination target behavior model based on the living environment information including the behavior history information and the surrounding situation information of the diagnosing doctor, and compares it with the comparison target behavior model when a diagnostic error has occurred, thereby predicting the occurrence of a diagnostic error in the diagnostic result of the diagnosing doctor. Therefore, it is possible to suppress an error in the diagnosis by the diagnosing doctor.

[0051] In the above embodiment, the pre-diagnosis situation DB 151 is generated by converting the pre-diagnosis situation of the diagnosing doctor into a database, but it may also be generated by converting and including the pre-diagnosis situations of doctors other than the diagnosing doctor into a database. In this case, a doctor ID may be assigned to each doctor whose pre-diagnosis situation is obtained to generate the pre-diagnosis situation DB 151, and the information classifying the doctors may be used to calculate the similarity of the action model.

[0052] (Second Embodiment) Subsequently, the second embodiment will be described. FIG. 4 is a block diagram showing an example of the configuration of the medical information processing apparatus 200 according to the second embodiment. The medical information processing apparatus 200 according to the second embodiment mainly differs from that of the first embodiment in that the processing circuit 140 includes a detection function 145 and an extraction function 146, and a diagnosis error detection model 153 and a diagnosis error information DB 154 are stored in the memory 150. Other points are common to the first embodiment. Hereinafter, the medical information processing apparatus 200 according to the second embodiment will be described centering on the differences from the first embodiment.

[0053] In the medical information processing apparatus 200 according to the first embodiment, the presence or absence of a diagnosis error in the diagnosis result is determined by a conference of the judging doctor group, and the determination result information of the diagnosis error is transmitted to the medical information processing apparatus. On the other hand, the medical information processing apparatus according to the second embodiment detects the occurrence of a diagnosis error by the detection function 145.

[0054] The detection function 145 generates, for example, a diagnosis error detection model 153 that outputs the presence or absence of a diagnosis error in the diagnosis result based on the pre-diagnosis situation stored in the pre-diagnosis situation DB 151 acquired by the acquisition function 141 and stored in the memory 150. The detection function 145 detects the occurrence or absence of a diagnosis error in the diagnosis result based on the generated diagnosis error detection model 153. The diagnosis error detection model 153 is an example of the first learned model.

[0055] When the detection function 145 is the confirmed determination result information obtained based on the determination result information transmitted by the conference server 50, for example, it assigns a determination result label corresponding to the determination result information to the pre-diagnosis situation included in the pre-diagnosis situation DB 151. When the obtained confirmed determination result information is error-containing information, for example, the detection function 145 assigns an error-present label as the determination result label to the pre-diagnosis situation of the diagnosing doctor before the diagnosis. When the obtained confirmed determination result information is error-free information, for example, the detection function 145 assigns an error-absent label as the determination result label to the pre-diagnosis situation of the diagnosing doctor before the diagnosis.

[0056] After the detection function 145 assigns a determination result label to the pre-diagnosis situation, it generates a diagnostic error detection model 153 using a number of pre-diagnosis situations with the assigned determination result label included in the pre-diagnosis situation DB 151. The detection function 145 estimates the pre-diagnosis situation at the time of error occurrence using the generated diagnostic error detection model 153.

[0057] FIG. 5 is a diagram showing an example of the content of the diagnostic error detection model 153. The diagnostic error detection model 153 is generated using the pre-diagnosis situation with the assigned determination result label as input data. The diagnostic error detection model 153 includes, for example, an input layer for inputting the input data, an output layer for outputting the output data, and an intermediate layer (hidden layer) between the input layer and the output layer. The intermediate layer has, for example, a multi-layer neural network connecting the input layer and the output layer. As the diagnostic error detection model 153, for example, a transformer, BERT, TFIDF (Term Frequency - Inverse Document Frequency), etc. may be used.

[0058] The diagnosis error detection model 153 is trained using teacher data with image information, audio information, biological information, etc. included in each of a large number of pre-diagnosis situations with assigned judgment result labels as input data and the presence or absence of a diagnosis error as output data. The diagnosis error detection model 153 may be generated by another device such as the conference server 50 or the medical information processing device 200, transmitted to the medical information processing device 200, and acquired by the detection function 145.

[0059] When the detection function 145 acquires the pre-determined judgment result information (hereinafter referred to as pre-determined judgment result information) transmitted by the conference server 50, it identifies the pre-diagnosis situation corresponding to the pre-determined judgment result information from the pre-diagnosis situation DB 151. The determined pre-diagnosis situation does not have an assigned judgment result label. The detection function 145 acquires the presence or absence of a diagnosis error, which is the output data when the identified pre-diagnosis situation is input to the diagnosis error detection model 153, as model judgment result information. The detection function 145 acquires, for example, error-positive information or error-negative information from the output data output by the diagnosis error detection model 153 as model judgment result information.

[0060] When the model judgment result information acquired by the detection function 145 using the diagnosis error detection model 153 is different from the pre-determined judgment result transmitted by the conference server 50, the detection function 145 transmits the difference information to the conference server 50. The difference information is information indicating that the model judgment result information and the pre-determined judgment result are different. When the conference server 50 receives the difference information, the judging doctor group reconsiders the diagnosis result, determines the presence or absence of a final diagnosis error, and transmits the judgment result to the medical information processing device 200 as the confirmed judgment result information.

[0061] When the model judgment result information and the pre-determined judgment result match, the detection function 145 generates the judgment result information as the confirmed judgment result information. When the confirmed judgment result information is error-positive information, the detection function 145 detects a diagnosis error. The detection function 145 is an example of a detection unit.

[0062] The detection function 145 generates a diagnostic error information database 154 that includes a pre-diagnosis situation with an error label (hereinafter referred to as a pre-diagnosis situation at the time of error occurrence). The detection function 145 stores the generated diagnostic error information database 154 in the memory 150. The pre-diagnosis situation at the time of error occurrence included in the diagnostic error information database 154 may be singular or plural. The pre-diagnosis situation at the time of error occurrence is an example of diagnostic error information.

[0063] The extraction function 146 extracts the pre-diagnosis situation at the time of error occurrence when a diagnostic error occurred in the past. The extraction function 146, for example, reads the diagnostic error information database 154 stored in the memory 150 and extracts the pre-diagnosis situation at the time of error occurrence when the pre-diagnosis situation was acquired by the acquisition function 141 based on the living environment information transmitted by the living environment sensor 10. The extraction function 146 is an example of an extraction unit.

[0064] The prediction function 144 calculates the similarity between the pre-diagnosis situation (hereinafter referred to as the determination target pre-diagnosis situation) acquired by the acquisition function 141 and the pre-diagnosis situation at the time of error occurrence (hereinafter referred to as the comparison target pre-diagnosis situation) extracted by the extraction function 146. The prediction function 144 predicts that there is a possibility of a diagnostic error when the calculated similarity between the determination target pre-diagnosis situation and the comparison target pre-diagnosis situation is equal to or greater than a predetermined threshold value.

[0065] Subsequently, the processing in the medical information processing apparatus 200 of the second embodiment will be described. FIGS. 6 to 8 are all flowcharts showing an example of the processing of the medical information processing apparatus 200 of the second embodiment. The medical information processing apparatus 200 of the second embodiment will describe the processing of generating a diagnostic error detection model 153, the processing of detecting a diagnostic error in past diagnoses, and the processing of predicting the possibility of a diagnostic error.

[0066] First, the process of generating the diagnostic error detection model 153 will be described. After the medical information processing device 200 receives the living environment information from the living environment sensor 10 and the determination result information from the conference server 50, it generates the diagnostic error detection model 153 or predicts the possibility of a diagnostic error. When executing the process of generating the diagnostic error detection model 153, the medical information processing device 200 of the second embodiment executes the processes from step S101 to step S111 in FIG. 3 according to the same procedure as the first embodiment shown in FIG. 3, and then executes the process shown in FIG. 6.

[0067] After the process of step S111 is completed, when the detection function 145 acquires the confirmed determination result information based on the determination result information transmitted by the conference server 50, it assigns a determination result label to the pre-diagnosis situation (step S201). Subsequently, the detection function 145 reads out a plurality of pre-diagnosis situations including the pre-diagnosis situation to which a new determination result label has been assigned (step S203).

[0068] Subsequently, the detection function 145 generates the diagnostic error detection model 153 based on the read pre-diagnosis situation (step S205). Subsequently, the detection function 145 stores the generated diagnostic error detection model 153 in the memory 150 (step S207). Thus, the medical information processing device 200 ends the process shown in FIG. 6.

[0069] Subsequently, the process of detecting a diagnostic error in past diagnoses will be described. FIG. 7 shows the flow of the process of detecting a diagnostic error in past diagnoses. The medical information processing device 100 first acquires, in the detection function 145, the pre-confirmed determination result information (hereinafter referred to as pre-confirmed determination result information) transmitted by the conference server 50 (step S221).

[0070] Subsequently, the detection function 145 identifies, as input data from among the pre-diagnosis situation DB 151, the pre-diagnosis situation corresponding to the acquired pre-determination result information (step S223). Subsequently, the detection function 145 acquires the model determination result information output by inputting the identified pre-diagnosis situation into the diagnosis error detection model 153 (step S223).

[0071] Subsequently, the detection function 145 compares the pre-determination result information transmitted by the conference server 50 with the model determination result information acquired by the detection function 145, and determines whether the pre-determination result information and the model determination result information are different (step S225). If it is determined that the pre-determination result information and the model determination result information are different, the detection function 145 generates difference information and transmits it to the conference server 50 (step S229).

[0072] When the conference server 50 receives the difference information, the judging medical staff group conducts a conference again regarding the presence or absence of a diagnosis error, and generates determination result information that becomes the confirmed determination result information. The generated confirmed determination result information is transmitted from the conference server 50 to the medical information processing apparatus 200.

[0073] The medical information processing apparatus 200 acquires, by means of the detection function 145, the confirmed determination result information transmitted by the conference server 50 (step S231). In step S227, if it is determined that the pre-determination result information and the model determination result information are not different (match), the detection function 145 generates the determination result information as the confirmed determination result information (step S233).

[0074] The detection function 145 determines the diagnosis result based on the acquired or generated confirmed determination result (step S235). The detection function 145 detects a diagnosis error when the confirmed determination result information is error-containing information. Subsequently, the detection function 145 determines whether a diagnosis error has been detected (step S237).

[0075] When it is determined that a diagnostic error has been detected, the detection function 145 generates a diagnostic error information DB 154 including a pre-diagnosis situation with an error label (hereinafter, the pre-diagnosis situation at the time of error occurrence) (step S239). The medical information processing apparatus 200 ends the process shown in FIG. 7. When it is determined by the detection function 145 that a diagnostic error has been detected, the medical information processing apparatus 200 ends the process shown in FIG. 7 as it is.

[0076] Subsequently, a process of predicting the possibility of a diagnostic error will be described. FIG. 8 shows the flow of the process of predicting the possibility of a diagnostic error. First, the acquisition function 141 of the medical information processing apparatus 100 acquires the pre-diagnosis situation of the diagnosing doctor based on the living environment information transmitted by the living environment sensor 10 and received by the communication interface 110 (step S251).

[0077] Subsequently, the extraction function 146 reads and extracts the pre-diagnosis situation at the time of error occurrence included in the diagnostic error information DB 154 stored in the memory 150 (step S253). The prediction function 144 calculates the similarity between the pre-diagnosis situation acquired by the extraction function 146 and the extracted pre-diagnosis situation at the time of error occurrence (step S255).

[0078] Subsequently, the prediction function 144 determines whether or not the calculated similarity is equal to or greater than a predetermined threshold (step S257). When it is determined that the calculated similarity is equal to or greater than the predetermined threshold, the prediction function 144 predicts that there is a possibility of a diagnostic error (step S259). When it is determined that the calculated similarity is less than the predetermined threshold, the prediction function 144 predicts that there is no possibility of a diagnostic error (step S261).

[0079] Subsequently, the prediction function 144 transmits the prediction result of the possibility of a diagnostic error to the conference server 50 (step S263). Thus, the medical information processing apparatus 200 ends the process shown in FIG. 8. The conference server 50 provides the transmitted prediction result to the judging doctor group. The judging doctor group determines the presence or absence of a diagnostic error in the diagnostic result of the diagnosis by the diagnosing doctor with reference to the provided prediction result.

[0080] The prediction function 144 determines whether there is a possibility of a diagnostic error based on whether the similarity between the pre-diagnosis situation and the pre-diagnosis situation at the time of gill generation is equal to or greater than a threshold value. However, the prediction function 144 may determine the possibility of a diagnostic error in other manners. Further, the prediction function 144 may determine the possibility of a diagnostic error as a ratio. In this case, for example, it may be predicted that the higher the similarity between the pre-diagnosis situation and the pre-diagnosis situation at the time of gill generation, the higher the possibility (ratio) of a diagnostic error.

[0081] The medical information processing apparatus 200 of the second embodiment has the same operational effects as the medical information processing apparatus 100 of the first embodiment. Further, the medical information processing apparatus 200 of the second embodiment includes a detection function 145. For this reason, since it is possible to determine whether or not there is a diagnostic error by the judging medical staff group and the diagnostic error detection model 153, the accuracy of determining a diagnostic error can be improved.

[0082] In the second embodiment, the detection function 145 is provided in the medical information processing apparatus 200. However, a function corresponding to the detection function may be provided in another apparatus, for example, the conference server 50. In this case, the medical information processing apparatus 200 may be configured to be able to acquire the confirmed determination result information generated by the conference server 50.

[0083] (Third Embodiment) Subsequently, the third embodiment will be described. FIG. 9 is a block diagram showing an example of the configuration of a medical information processing apparatus 300 according to the third embodiment. The medical information processing apparatus 300 according to the third embodiment mainly differs from the second embodiment in that the processing circuit 140 includes a speculation function 147 and a notification function 148. Other points are common to the second embodiment. Hereinafter, the medical information processing apparatus 300 according to the third embodiment will be described centering on the differences from the second embodiment.

[0084] The speculation function 147 speculates on characteristic pre-diagnosis situations (hereinafter referred to as characteristic situations) based on the pre-diagnosis situations at the time of error occurrence extracted by the extraction function 146. The characteristic situation is a characteristic pre-diagnosis situation that contributed to the occurrence of the diagnostic error. The speculation function 147 extracts, for example, pre-diagnosis situations common to a plurality of pre-diagnosis situations at the time of error occurrence. The speculation function 147 speculates that the pre-diagnosis situation common to the extracted plurality of pre-diagnosis situations at the time of error occurrence is the characteristic situation. The storage function 143 stores the characteristic situation speculated by the speculation function 147 in the memory 150 to generate the characteristic information DB 155. The speculation function 147 is an example of a speculation unit.

[0085] Based on the pre-diagnosis situation acquired by the acquisition function 141 and the characteristic pre-diagnosis situation included in the characteristic information DB 155 stored in the memory 150, the speculation function 147 speculates on the occurrence of a similar diagnostic error occurrence situation similar to the diagnostic error occurrence situation. The prediction function 144 predicts the occurrence of a diagnostic error based on the similar diagnostic error occurrence situation speculated by the speculation function 147. For example, the prediction function 144 calculates the similarity between the diagnosis target pre-diagnosis situation and the characteristic situation, and predicts that there may be a diagnostic error in the diagnosis by the diagnostician when the similarity is equal to or greater than the threshold value.

[0086] The notification function 148 notifies by, for example, displaying various information related to the occurrence of a diagnostic error on the display 130. The notification function 148 notifies prediction information according to the prediction result, for example, when the occurrence of a diagnostic error is predicted by the prediction function 144. The notification function 148 notifies speculation information according to the speculation result, for example, when the occurrence of the diagnostic error occurrence situation is speculated by the speculation function 147. The notification function 148 notifies the occurrence of a diagnostic error, for example, when the occurrence of a diagnostic error is detected by the detection function 145. The notification function 148 is an example of a notification unit.

[0087] The speculation function 147 uses the characteristic diagnostic pre - situation included in the characteristic information DB155 stored in the memory 150 to speculate on the occurrence of a similar diagnostic error occurrence situation similar to the situation at the time of the diagnostic error. In contrast, the speculation function 147 may speculate on the occurrence of a similar diagnostic error occurrence situation based on the error occurrence situation speculation model 156 that outputs speculation information regarding the occurrence of a similar diagnostic error occurrence situation based on the pre - diagnostic situation information. The error occurrence situation speculation model 156 is an example of a second pre - trained model.

[0088] FIG. 10 is a diagram showing an example of the content of the error occurrence situation speculation model 156. The error occurrence situation speculation model 156 is generated using the pre - diagnostic situation with a determination result label as input data. The error occurrence situation speculation model 156 includes, for example, an input layer for inputting input data, an output layer for outputting output data, and an intermediate layer (hidden layer) between the input layer and the output layer. The intermediate layer has, for example, a multi - layer neural network connecting the input layer and the output layer. As the error occurrence situation speculation model 156, for example, a transformer, BERT, TFIDF, etc. may be used.

[0089] The error occurrence situation speculation model 156 is learned using teacher data with image information, voice information, biological information, etc. included in each of a large number of pre - diagnostic situations with a determination result label as input data and the diagnostic error occurrence situation as output data. The error occurrence situation speculation model 156 may be generated by another device such as the conference server 50 or the medical information processing device 200, transmitted to the medical information processing device 200, and acquired by the speculation function 147.

[0090] The medical information processing device 300 of the third embodiment has the same effects as the medical information processing device 200 of the above - described second embodiment. Furthermore, the medical information processing device 300 of the third embodiment focuses on the characteristic situation among the pre - diagnostic situations when a diagnostic error occurs and uses it for predicting the occurrence of the diagnostic error. Therefore, it is possible to more accurately predict the occurrence of a diagnostic error in the diagnostic result of the diagnosing doctor.

[0091] In each of the above embodiments, the pre-diagnosis situation used is the pre-diagnosis situation of the diagnosing doctor himself / herself, but pre-diagnosis situations of others, for example, pre-diagnosis situations of doctors other than the doctor himself / herself or pre-diagnosis situations of patients receiving diagnoses, may also be used. In this case, each model such as the behavior model based on the pre-diagnosis situation may be generated using the pre-diagnosis situation labeled with whether it is a diagnosing doctor or not.

[0092] According to at least one of the embodiments described above, the medical information processing apparatus includes an acquisition unit that acquires behavior history information regarding the doctor's behavior history before diagnosis and peripheral situation information regarding the situation around the doctor, diagnosis pre-situation information regarding a pre-diagnosis situation including at least one of the behavior history or the peripheral situation, and a prediction unit that predicts the occurrence of a diagnostic error of the doctor based on diagnosis error information regarding past diagnostic errors and the situation at the time of occurrence of the diagnostic error. By having these components, it is possible to suppress diagnostic errors.

[0093] Although several embodiments have been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope and gist of the invention, and are also included in the invention described in the claims and its equivalent scope.

Explanation of Reference Numerals

[0094] 10 Living environment sensor 12 Surveillance camera 14 Voice recorder 16 Wearable sensor 50 Conference server 100, 200, 300 Medical information processing apparatus 110 Communication interface 120 Input interface 130 Display 140 Processing circuit 141 Acquisition function 142 Generation function 143 Storage function 144 Prediction function 145 Detection function 146 Extraction function 147 Inference function 148 Notification function 150 Memory 151 Pre-diagnosis situation DB 152 Diagnosis error prediction model 153 Diagnosis error detection model 154 Diagnosis error information DB 155 Characteristic information DB 156 Error occurrence situation inference model

Claims

1. An acquisition unit that acquires life environment information including action history information regarding the action history of a doctor before diagnosis and peripheral situation information regarding the situation around the doctor; A prediction unit that predicts the occurrence of a diagnostic error of the doctor based on the pre-diagnosis situation information regarding the pre-diagnosis situation including the life environment information before the doctor makes a diagnosis and the diagnostic error information regarding past diagnostic errors and the situation at the time of occurrence of the diagnostic error; A medical information processing device.

2. The prediction unit acquires the diagnostic error information from a storage unit that stores the diagnostic error information. The medical information processing device according to Claim 1.

3. The past diagnostic errors included in the diagnostic error information and the situations at the time of occurrence of the diagnostic errors are linked to each other. The medical information processing device according to Claim 1.

4. The life environment information includes the life environment information of at least one of the doctor or the patient. The medical information processing device according to Claim 1.

5. It further includes an accumulation unit that stores and accumulates the acquired life environment information in the storage unit. The medical information processing device according to Claim 2.

6. A detection unit that detects the occurrence of a diagnostic error; It further includes an extraction unit that extracts the pre-diagnosis situation at the time of error occurrence when the diagnostic error occurred in the past. The medical information processing device according to Claim 5.

7. It further includes an accumulation unit that stores and accumulates the extracted pre-diagnosis situation at the time of error occurrence in the storage unit. The medical information processing device according to Claim 6.

8. It further includes a speculation unit that speculates a characteristic pre-diagnosis situation that is a characteristic pre-diagnosis situation contributing to the occurrence of the diagnostic error based on the pre-diagnosis situation at the time of error occurrence extracted by the extraction unit, The accumulation unit stores and accumulates the speculated characteristic pre-diagnosis situation in the storage unit. The medical information processing device according to Claim 7.

9. The speculation unit speculates the occurrence of a similar diagnostic error occurrence situation similar to the diagnostic error occurrence situation based on the pre-diagnosis situation based on the life environment information acquired by the acquisition unit and the characteristic pre-diagnosis situation stored in the storage unit. The medical information processing device according to Claim 8.

10. The detection unit detects the presence or absence of the occurrence of the diagnostic error based on a first learned model that outputs the presence or absence of the occurrence of the diagnostic error based on the pre-diagnosis situation. The medical information processing device according to Claim 6.

11. When the occurrence of the diagnostic error is predicted by the prediction unit, further comprising a notification unit that notifies prediction information according to the result of the prediction. The medical information processing apparatus according to claim 1.

12. When the occurrence of the situation at the time of the diagnostic error is inferred by the inference unit, further comprising a notification unit that notifies inference information according to the result of the inference. The medical information processing apparatus according to claim 9.

13. Based on the second learned model that outputs inference information regarding the occurrence of a similar diagnostic error occurrence situation based on the pre-diagnosis situation information, the inference unit infers the occurrence of the similar diagnostic error occurrence situation. The medical information processing apparatus according to claim 12.

14. Based on the bone model generated based on the pre-diagnosis situation information, the prediction unit predicts the occurrence of the diagnostic error of the doctor. The medical information processing apparatus according to claim 1.

15. When the occurrence of the diagnostic error is detected by the detection unit, further comprising a notification unit that notifies the occurrence of the diagnostic error. The medical information processing apparatus according to claim 6.

16. A computer acquires living environment information including action history information regarding the doctor's action history before diagnosis and peripheral situation information regarding the situation around the doctor, and predicts the occurrence of the doctor's diagnostic error based on the pre-diagnosis situation information regarding the pre-diagnosis situation including the living environment information before the doctor makes a diagnosis, and the diagnostic error information regarding the past diagnostic error and the diagnostic error occurrence situation at the time of the occurrence of the diagnostic error. Medical information processing method.

17. Causing a computer to acquire living environment information including action history information regarding the doctor's action history before diagnosis and peripheral situation information regarding the situation around the doctor, and to predict the occurrence of the doctor's diagnostic error based on the pre-diagnosis situation information regarding the pre-diagnosis situation including the living environment information before the doctor makes a diagnosis, and the diagnostic error information regarding the past diagnostic error and the diagnostic error occurrence situation at the time of the occurrence of the diagnostic error. Program.

Images