Selection assistance device, selection assistance method, and recording medium

Abstract

This selection assistance device comprises an acquisition unit, a conversion unit, a generation unit, and an output unit. The acquisition unit acquires a sentence indicating a selection criterion for clinical trial patients. The conversion unit converts the sentence indicating the selection criterion into a structured selection criterion. The generation unit generates a reference instruction indicating information to be referred to in the selection of clinical trial patients on the basis of the structured selection criterion. The output unit outputs the generated reference instruction. The use of the reference instruction generated on the basis of the structured selection criterion allows for appropriate decision making in screening of clinical trial patients.

Description

Selection support device, selection support method, and recording medium

[0001] The present disclosure relates to a selection support device and the like.

[0002] In a clinical trial of a pharmaceutical product, for example, a person in charge of the clinical trial entrusted by a pharmaceutical company sends the selection criteria for clinical trial patients to a medical institution and inquires about information on patients who are suitable for the clinical trial. The medical institution's person in charge, upon receiving the inquiry, checks the medical records to extract patients who meet the selection criteria for clinical trial patients. The medical institution's person in charge then responds to the inquiry. The medical institution's person in charge of extracting patients who meet the selection criteria determines, for example, whether the selection criteria match the details written in the medical records. The person in charge of extracting patients who meet the selection criteria then extracts patients whose selection criteria match the details written in the electronic medical records as candidate clinical trial patients. In addition, a system that supports the extraction may be used to extract candidate clinical trial patients.

[0003] The clinical trial candidate extraction device of Patent Document 1 extracts, as clinical trial candidates, patients who are likely to participate in the clinical trial from among patients currently being treated for the disease targeted by the investigational drug.

[0004] JP 2014-194595 A

[0005] With the technology described in Patent Document 1, it may be difficult to determine where to check information for each patient.

[0006] In order to solve the above-mentioned problems, the present disclosure aims to provide a selection support device, etc. that makes it easy to identify information to refer to when selecting clinical trial patients.

[0007] In order to solve the above problems, the selection support device disclosed herein comprises an acquisition means for acquiring text indicating the selection criteria for clinical trial patients, a conversion means for converting the text indicating the selection criteria into structured selection criteria, a generation means for generating reference instructions indicating information to be referred to when selecting clinical trial patients based on the structured selection criteria, and an output means for outputting the generated reference instructions.

[0008] The selection support method disclosed herein acquires text indicating the selection criteria for clinical trial patients, converts the text indicating the selection criteria into structured selection criteria, generates reference instructions indicating information to be referenced in selecting clinical trial patients based on the structured selection criteria, and outputs the generated reference instructions.

[0009] The recording medium of the present disclosure non-temporarily records a selection support program that causes a computer to execute the following processes: acquiring text indicating selection criteria for clinical trial patients; converting the text indicating the selection criteria into structured selection criteria; generating reference instructions indicating information to be referenced in selecting clinical trial patients based on the structured selection criteria; and outputting the generated reference instructions.

[0010] The present disclosure makes it easier to identify information to refer to when selecting clinical trial patients.

[0011] FIG. 1 is a diagram illustrating an example of the configuration of a clinical trial support system according to the present disclosure. FIG. 2 is a diagram illustrating an example of the configuration of a selection support device according to the present disclosure. FIG. 3 is a diagram illustrating an example of the configuration of inclusion criteria text according to the present disclosure. FIG. 4 is a diagram illustrating an example of the configuration of exclusion criteria text according to the present disclosure. FIG. 5 is a diagram illustrating an example of the configuration of structured inclusion criteria text according to the present disclosure. FIG. 6 is a diagram illustrating an example of the configuration of structured exclusion criteria text according to the present disclosure. FIG. 7 is a diagram illustrating an example of a display screen according to the present disclosure. FIG. 8 is a diagram illustrating an example of a display screen according to the present disclosure. FIG. 9 is a diagram illustrating an example of the operation flow of a selection support device according to the present disclosure. FIG. 10 is a diagram illustrating an example of the hardware configuration of a selection support device according to the present disclosure.

[0012] Embodiments of the present disclosure will be described in detail with reference to the drawings. FIG. 1 is a diagram showing an example of the configuration of a clinical trial support system. The clinical trial support system includes a selection support device 10 and a terminal device 20. The selection support device 10 is connected to the terminal device 20, for example, via a network. There may be multiple terminal devices 20. The number of terminal devices 20 can be set as appropriate.

[0013] A clinical trial support system is, for example, a system that outputs reference instructions indicating information to be referenced when selecting clinical trial patients. The information to be referenced when selecting clinical trial patients is, for example, information that needs to be confirmed when determining whether a patient meets the selection criteria for clinical trial patients. A clinical trial support system is used, for example, when a clinical trial to verify the effectiveness of a new drug is planned, to extract patients to be selected as clinical trial patients. A clinical trial support system may also be used to estimate the number of clinical trial patients during the preparation stage of a clinical trial protocol. The uses of a clinical trial support system are not limited to the above. A clinical trial is, for example, a clinical trial conducted to obtain legal approval for the manufacture and sale of pharmaceuticals or medical devices. A clinical trial patient is, for example, a patient who is a subject of a clinical trial. The reference instructions are instructions regarding information that needs to be confirmed when determining whether a patient is a clinical trial patient. The reference instructions include, for example, the location where the information to be referenced is described and the content of the information to be referenced. The reference instructions are, for example, a prompt used as input for a large-scale language model. The reference instructions are not limited to prompts.

[0014] The information to be referenced in selecting clinical trial patients is, for example, information that must be confirmed in detail to ensure that it meets the selection criteria when determining whether a patient is a candidate for a clinical trial patient. The information to be referenced in selecting clinical trial patients is, for example, information about the patient's treatment. For example, a person who extracts patients who meet the selection criteria for clinical trial patients, such as a doctor, will check the information about the treatment based on the reference instructions to determine whether each patient meets the selection criteria.

[0015] The information regarding treatment is, for example, records related to the patient's treatment. The information regarding treatment is, for example, information recorded in a medical record and test data. The information recorded in a medical record is, for example, information recorded in an electronic medical record. The information regarding treatment may be, for example, either information recorded in a medical record or test data. The information regarding treatment may also be information included in medical accounting data or medical receipts. A medical receipt is, for example, a statement of medical fees that a medical institution bills an insurer. A medical institution may include a pharmacy that prescribes medication. The information regarding treatment may also be information included in medical literature. Medical literature may include guidelines regarding treatment. The information regarding treatment is not limited to the above.

[0016] The selection criteria are, for example, inclusion criteria and exclusion criteria for patients in a clinical trial. The selection criteria may be either inclusion criteria or exclusion criteria for patients in a clinical trial. The inclusion criteria are, for example, information indicating the conditions for patients to be eligible for the clinical trial. The conditions for patients to be eligible for the clinical trial are indicated using the attributes and characteristics of patients suitable for the clinical trial. The exclusion criteria are information indicating the conditions for excluding patients from the clinical trial. That is, the exclusion criteria are, for example, information indicating the conditions for patients not to be selected as clinical trial patients. The exclusion criteria are indicated using the attributes and characteristics of patients not suitable for the clinical trial. The patient attributes are, for example, a patient condition that does not change with treatment. The patient attributes are, for example, information on one or more of the patient's age, sex, weight, height, and race. The patient attributes are not limited to the above. The patient characteristics are, for example, information on one or more of the treatments, surgeries, tests, medications, test results, and follow-up results performed on the patient. The patient characteristics are not limited to the above.

[0017] Extraction of patients who meet the selection criteria for clinical trial patients is performed, for example, by a person in charge at a medical institution. In this case, the selection support device 10 can be accessed, for example, by a person in charge at the medical institution. The person in charge at the medical institution is, for example, one or both of a person belonging to the medical institution and a person in charge of an institution outsourced by the medical institution. An institution outsourced by a medical institution to handle clinical trials is, for example, an SMO (Site Management Organization). Also, a person in charge at an institution outsourced by a hospital to handle clinical trials is, for example, a CRC (Clinical Research Coordinator). The person in charge at a medical institution is not limited to the above.

[0018] For example, a person in charge at a medical institution receives an inquiry from a person in charge at a pharmaceutical company about information on patients who fit the selection criteria for patients to be included in a clinical trial. For example, a person in charge at an institution entrusted with a clinical trial by a pharmaceutical company conducting a clinical trial of a new drug sends the selection criteria to the person in charge at the medical institution and inquires about information on patients who fit the selection criteria.

[0019] The person in charge at a pharmaceutical company may be, for example, a person in charge at an institution that has been entrusted with a clinical trial by a pharmaceutical company that is conducting a clinical trial of a new drug. The institution that has been entrusted with a clinical trial by a pharmaceutical company may be, for example, a CRO (Contract Research Organization). The person in charge at the institution that has been entrusted with a clinical trial by a pharmaceutical company may be, for example, a CRA (Clinical Research Associate). The CRA may be an employee of the pharmaceutical company.

[0020] An inquiry about information on patients who meet the selection criteria for patients to be included in a clinical trial is made, for example, when a CRA prepares a clinical trial protocol. An inquiry about information on patients who meet the selection criteria for patients to be included in a clinical trial may also be made when the clinical trial is being carried out in accordance with the clinical trial protocol. The clinical trial protocol is used, for example, to explain the contents of the clinical trial to and negotiate with medical institutions, and to submit notifications to related institutions. The uses of the clinical trial protocol are not limited to the above. Furthermore, the timing for an inquiry about information on patients who meet the selection criteria is not limited to the stage of preparing the clinical trial protocol.

[0021] Here, a specific example of the configuration of the selection support device 10 will be described. Fig. 2 is a diagram showing an example of the configuration of the selection support device 10. The selection support device 10 basically includes an acquisition unit 11, a conversion unit 12, a generation unit 14, and an output unit 15. The selection support device 10 may also include, for example, an identification unit 13 and a storage unit 16.

[0022] The acquisition unit 11 acquires a sentence indicating the selection criteria for clinical trial patients. The acquisition unit 11 acquires the sentence indicating the selection criteria, for example, from the terminal device 20. The sentence indicating the selection criteria is input to the terminal device 20, for example, by a CRC. The person who inputs the sentence indicating the selection criteria to the terminal device 20 is not limited to a CRC. The sentence indicating the selection criteria is also created, for example, by a CRA. Then, the sentence indicating the selection criteria is sent, for example, from the CRA to the CRC. The person who creates the sentence indicating the selection criteria is not limited to a CRA.

[0023] Figures 3 and 4 are examples of sentences indicating the selection criteria for clinical trial patients. The sentences indicating the selection criteria are, for example, sentences describing each of the criteria included in the selection criteria. Figures 3 and 4 are examples of sentences indicating the selection criteria for a clinical trial of a drug used to treat breast cancer. Figure 3 is an example of a document describing the inclusion criteria, which are part of the selection criteria. The example of the inclusion criteria in Figure 3 includes 15 sentences indicating the inclusion criteria. Each sentence in the inclusion criteria indicates the conditions for patients to be selected as clinical trial patients. For example, the sentence "Be 18 years of age or older at the time of screening" indicates that one of the conditions for selection as a clinical trial patient is being 18 years of age or older at the time of screening. Figure 4 is also an example of sentences indicating the exclusion criteria for a clinical trial. The example of the exclusion criteria in Figure 4 includes 19 sentences indicating the exclusion criteria. Each sentence in the exclusion criteria indicates the conditions for patients not to be selected as clinical trial patients. For example, the sentence "Having poorly controlled or significant heart disease" indicates that one of the conditions for not being selected as a clinical trial patient is having poorly controlled or significant heart disease. In the example sentences showing the selection criteria in Figures 3 and 4, for example, patients who meet all of the inclusion criteria shown in 15 sentences and do not meet any of the exclusion criteria shown in 19 documents are selected as clinical trial patients. In addition, the selection of clinical trial patients may be performed using criteria in which some of the selection criteria are relaxed.

[0024] The acquisition unit 11 may acquire an item for generating a reference indication. The item for generating a reference indication is information that specifies the content of the criterion that is the target of the reference indication, among the selection criteria. The information that specifies the content of the criterion that is the target of the reference indication may be a superordinate concept of the criterion for generating a reference indication. For example, if the target for generating a reference indication is a criterion related to "lung cancer," the item would be "disease name." The items are not limited to the above. The identification unit 13 acquires the item for generating a reference indication from, for example, the terminal device 20. The item for generating a reference indication is input, for example, to the terminal device 20 by the operation of the person to be extracted.

[0025] The conversion unit 12 converts the sentence indicating the selection criteria into structured selection criteria. The structured selection criteria is, for example, data that associates items included in the sentence indicating the selection criteria with the conditions for each item. The items, for example, indicate the content of each criterion included in the selection criteria. For example, if the criteria identify the name of the disease that a patient is suffering from, the item is "disease name." The content of the criteria included in the items is not limited to the above. For example, the conversion unit 12 extracts the selection criteria items and the conditions for each item from the sentence indicating the selection criteria. Then, the conversion unit 12 converts the extracted selection criteria items into data that associates the extracted selection criteria items with the conditions for each item.

[0026] The conversion unit 12, for example, adds labels to terms included in the sentence indicating the selection criteria. The terms are, for example, terms related to the selection of clinical trial patients. The terms related to the selection of clinical trial patients are terms related to information that needs to be confirmed when selecting clinical trial patients. The information that needs to be confirmed when selecting clinical trial patients is, for example, information indicating the patient's condition that may affect the conduct and results of the clinical trial. The terms related to the selection of clinical trial patients are, for example, terms indicating one or more of the patient's condition, such as disease name, treatment content, drug name, test, test result, medical condition, and pre-existing condition. The terms related to the selection of clinical trial patients may include patient attributes. The patient attributes are information indicating a patient's condition that does not change with treatment. The patient attributes are, for example, one or more of the patient's age, gender, weight, height, race, and family medical history. The patient attributes are not limited to the above. Furthermore, the terms related to the selection of clinical trial patients are not limited to the above. The conversion unit 12, for example, converts the sentence indicating the selection criteria into structured selection criteria based on the added labels. The conversion unit 12 uses, for example, a labeling model to add labels to terms included in the sentence indicating the selection criteria. The labeling model is a machine learning model that uses natural language processing. For example, Word2vec can be used as the natural language processing. The natural language processing algorithm is not limited to the above. For example, the labeling model uses a dictionary in the medical field to break down the sentence indicating the selection criteria into terms through morphological analysis. Then, for example, the labeling model adds labels to each term using the dictionary. For example, the labeling model is generated by learning the relationship between the sentence indicating the selection criteria, the dictionary, and the labels to be added. For example, the labeling model is generated by deep learning using a neural network. For example, the labeling model is generated in a system external to the selection support device 10.

[0027] The dictionary includes, for example, data associating medical terms with the contents of their labels. The contents of the labels are, for example, information indicating the classification of the terms. The contents of the labels may include, for example, the part of speech of the terms. The classification of the terms is, for example, information indicating the context in which the terms are used in the medical field, such as disease names, time periods, patients, guideline names, standards, drug names, and conditions to which the terms apply. The context in which the terms are used is, for example, examinations, medications, and treatments. The context in which the terms are used is not limited to the above. The classification of the terms is also not limited to the above. The labels may also include information regarding the inclusion relationships between terms. For example, the information regarding the inclusion relationships between terms is information indicating the hierarchical relationship between the meanings of the terms. For example, given the terms "lung cancer" and "non-small cell lung cancer," "lung cancer" is a term that includes "non-small cell lung cancer." In this case, "lung cancer" is a higher-ranking term and a lower-ranking term than "non-small cell lung cancer." The labels may also include information indicating the chronological order of the terms. For example, when adding a label to the sentence "After administering drug A, it was discontinued due to side effects and drug B was administered," the labeling model adds a label indicating that "drug A" and "administration" come before "drug B" and "administration" in the chronological order.

[0028] The conversion unit 12 generates structured selection criteria based on, for example, labeled selection criteria. Structuring refers to, for example, converting mutually related information contained in a sentence into a state in which the relationships between the information and the conditions for each item are clear. For example, when generating structured selection criteria, structuring refers to extracting items corresponding to each criterion contained in the selection criteria and the conditions for each item, and converting the sentence indicating the selection criteria into a state in which the relationships between the items and the conditions for each item are clear. For example, the conversion unit 12 identifies terms in the sentence of the selection criteria to which labels related to the items are attached. The relationship between the labels and the items is set, for example, as data in a table format. For example, the conversion unit 12 extracts terms in the sentence of the selection criteria to which labels related to the items are attached and the conditions for each item. For example, if there is a term labeled with a drug name, the conversion unit 12 extracts medication history as an item. Furthermore, the conversion unit 12 extracts, for example, "drug A," "side effects," and "discontinuation" as conditions corresponding to the item. After extracting the items and the conditions for each item, the conversion unit 12 generates data as structured selection criteria, in which terms to which the item labels have been added are associated with the conditions for each item.

[0029] Figures 5 and 6 are diagrams showing examples of structured selection criteria. Figure 5 shows an example of structured inclusion criteria among the structured selection criteria. Figure 6 shows an example of structured exclusion criteria among the structured selection criteria.

[0030] In the example of the inclusion criteria in Fig. 5, for example, "age," which is an item included in the inclusion criteria, is associated with the condition corresponding to the item, "18 years of age or older." In this case, one of the criteria included in the inclusion criteria is that the patient's age is 18 years of age or older.

[0031] In the example of the exclusion criteria in FIG. 6 , for example, the item "pre-existing condition" included in the exclusion condition is associated with the condition "autoimmune or inflammatory disease" corresponding to the item. In this case, one of the criteria included in the exclusion criteria is that the patient has an autoimmune or inflammatory disease in their past medical history. In other words, if a patient has an autoimmune or inflammatory disease in their past medical history, the patient is excluded from the list of candidates for clinical trial patients. Furthermore, examples of structured selection criteria are not limited to the examples in FIGS. 5 and 6 . For example, in the structured selection criteria, each item may be hierarchically organized. For example, in the item "medication history," the items "drug name," "medication period," and "dosage" may be associated below "medication history." Furthermore, the hierarchical structure may have three or more layers.

[0032] The conversion unit 12 may use a structured model to convert a sentence indicating the selection criteria into a structured selection criteria. The structured model, for example, receives a sentence indicating the selection criteria with labels and a format of the structured selection criteria as input, and outputs the structured selection criteria. The structured model, for example, is generated by learning the relationship between the sentence indicating the selection criteria with labels, the format of the structured selection criteria, and structured data. The labeled structured model may receive a sentence indicating the selection criteria as input, and output the structured selection criteria. In this case, the structured model is generated by learning the relationship between the sentence indicating the selection criteria with labels and the structured data. The structured model is generated, for example, by deep learning using a neural network. The structured model is generated, for example, in a system external to the selection support device 10.

[0033] The selection criteria format is information that defines how to divide the selection criteria into items when structuring the selection criteria. The selection criteria format is set, for example, for each disease to which the drug being investigated is administered. In clinical trials, for example, evaluation items are often defined for each disease. For this reason, for example, by using the items to be extracted as selection criteria as input together with the text indicating the selection criteria, it is possible to extract the conditions corresponding to the items from the text indicating the selection criteria.

[0034] A large-scale language model may be used as the structured model. Examples of the large-scale language model that may be used include Generative Pre-Training-2 (GPT-2), GPT-3, and GPT-4. Examples of the large-scale language model that may be used include Text-to-Text Transfer Transformer (T5), Bidirectional Encoder Representations from Transformers (BERT), Robustly optimized BERT approach (RoBERTa), and Efficiently Learning an Encoder that Classifies Token Replacements Accurately (ELECTRA). The structured model may be generated, for example, in a system external to the selection support device 10. The process of generating structured selection criteria using the structured model may be performed in a system external to the selection support device 10.

[0035] The identification unit 13, for example, identifies a target for which a reference indication is to be generated. The identification unit 13, for example, identifies a target for which a reference indication is to be generated based on the degree to which confirmation is necessary in the selection of clinical trial patients. For example, in the structuring process, it is desirable for a person in charge to thoroughly check items for which the accuracy of conversion to the structured selection criteria may be low and the conditions described in the items. A low accuracy of conversion to the structured selection criteria means, for example, that the compatibility between the terms included in the selection criteria and the terms in the dictionary is low, and there is a high possibility that the sentences in the selection criteria have not been correctly analyzed. In this case, the identification unit 13 identifies a target for which a reference indication is to be generated based on, for example, the reliability of the structuring process. The identification unit 13 identifies an item for which a reference indication is to be generated based on, for example, the reliability of the terms assigned to the items in the structured selection criteria. The identification unit 13 may also identify an item in the structured selection criteria for which the conditions are not clearly defined as a target for which a reference indication is to be generated.

[0036] The identification unit 13 calculates a score for each item included in the selection criteria based on, for example, the degree of match between the terms included in the selection criteria and the terms included in the dictionary. The score is, for example, an index for identifying an item for which a reference indication is to be generated. The score is, for example, an example of an index indicating the reliability of the structuring process. For example, as the reliability of the structuring process decreases, the need for the person in charge to check the information in detail increases. Therefore, the score can be an index indicating the need for the person in charge to check the information, for example, when selecting clinical trial patients. For example, the score is calculated so that the value decreases as the need for generating a reference indication increases. The degree of match of a term is, for example, the degree of match between a term included in a sentence indicating the selection criteria and a term included in a dictionary. For example, the higher the degree of match between a term included in a dictionary and a term included in the sentence of the selection criteria, the higher the value of match. Furthermore, the score for each item included in the selection criteria increases, for example, as the degree of match of the term increases. The identification unit 13 then identifies an item with a score below the standard as a target for which a reference indication is to be generated. The score criteria for identifying a target for generating a reference instruction are set based on, for example, the need to check the information in the electronic medical record in detail. For example, the score criteria are set so that, when the score is below the criteria, it is desirable to check the information in the electronic medical record in detail.

[0037] The identification unit 13 may identify items for which reference instructions are to be generated based on criteria set for each disease name or condition. The condition is, for example, the degree of progression of the disease. For example, the selection criteria items that are emphasized may change depending on the stage indicating the degree of progression of cancer. Even a slight difference in the selection criteria items that are emphasized may affect the results of the clinical trial. Therefore, for example, by establishing rules for the items for which reference instructions are to be generated for each disease name or condition, reference instructions can be generated for items corresponding to the disease name or condition. The identification unit 13 may, for example, generate reference instructions for items set for each disease name. The identification unit 13 may generate reference instructions for items set for each disease name and condition. The identification unit 13 may identify items for which reference instructions are to be generated based on the disease name extracted from the selection criteria. For example, if the selection criteria include "breast cancer," the identification unit 13 identifies items set for breast cancer as items for which reference instructions are to be generated.

[0038] The identification unit 13 may generate a reference indication based on the item for generating a reference indication acquired by the acquisition unit 11. The identification unit 13 acquires the item for generating a reference indication, for example, from the terminal device 20. The item for generating a reference indication is input to the terminal device 20 by an operation of the person to be extracted, for example.

[0039] The generation unit 14 generates a reference instruction for information to be referenced in selecting clinical trial patients based on the structured selection criteria. The reference instruction includes, for example, a location where the information to be referenced is described and the content to be referenced. The reference instruction includes, for example, a location in an electronic medical record where the information to be referenced is described and the content to be referenced.

[0040] The generation unit 14 generates a reference indication based on a set rule. The rule is set, for example, as table-format data. The generation unit 14 generates the reference indication by, for example, referring to a table that associates items included in structured selection criteria with locations of treatment-related data. The locations of treatment-related data are, for example, information indicating a document in which information corresponding to the selection criteria is described and the location of the information in the document. The generation unit 14 may also generate a prompt-format reference indication to be used as input for a large-scale language model that extracts information from treatment-related information.

[0041] The generator 14 may generate the reference indication using a generative model. The generative model takes items included in the selection criteria as input and outputs locations where information related to treatment is described. The generative model may take items included in the selection criteria and conditions associated with the items as input and output locations where information related to treatment is described. The generative model may be generated, for example, by learning the relationship between items included in the selection criteria and locations where information related to treatment is described. The generative model may also be generated, for example, by learning the relationship between items included in the selection criteria and conditions associated with the items and locations where information related to treatment is described.

[0042] When the identification unit 13 identifies an item for which a reference indication is to be generated, the generation unit 14 generates, for example, a reference indication for the item identified by the identification unit 13. The generation unit 14 may also generate a reference indication for a predetermined item. The predetermined item is set using an item of the structured selection criteria or a term included in the item. The generation unit 14, for example, determines an item of the structured selection criteria or a term included in the item. Then, the generation unit 14 generates a reference indication for the item determined to be a predetermined item. The predetermined item is, for example, an item that requires detailed confirmation. An item that requires detailed confirmation is, for example, an item for which, in selecting clinical trial patients, it may be difficult to accurately select clinical trial patients unless the contents recorded in the electronic medical record are confirmed in detail. For example, if the administration history of other drugs has a significant impact on the clinical trial, an item that describes conditions related to the administration history may be an item that requires detailed confirmation.

[0043] The output unit 15 outputs the reference indication generated by the generation unit 14. The output unit 15 outputs the reference indication to, for example, the terminal device 20. Furthermore, when the generation unit 14 generates a reference indication in the prompt format, the output unit 15 outputs the reference indication in the prompt format to be used as an input for a large-scale language model, for example.

[0044] The output unit 15 may output the structured selection criteria together with the reference indication. The output unit 15 may output the structured selection criteria instead of the reference indication. Furthermore, the output unit 15 may output the structured selection criteria when the structured selection criteria do not include an item for which a reference indication is to be generated. Furthermore, the output unit 15 may output a score associated with each item included in the structured selection criteria. The score is, for example, an index indicating the likelihood of converting a sentence indicating the selection criteria into structured structure data.

[0045] 7 is a diagram showing an example of an output screen for information to be referenced. The example display screen in FIG. 7 displays "Extraction of treatment information from doctor's findings (extraction of medication and dosage)" and "Extraction of drug efficacy information from image information (extraction of changes in cancer size and side effects)." A person in charge of extracting candidate clinical trial patients can, for example, refer to the example display screen in FIG. 7 to check the contents of the electronic medical record.

[0046] 8 and 9 are examples of a display screen displaying structured selection criteria. The example display screen of FIG. 8 is a screen displaying inclusion criteria from among the structured selection criteria. In the example display screen of FIG. 8, the criteria items included in the inclusion criteria are associated with the conditions for each item. Furthermore, the items and the conditions for each item are further associated with a score. The score is, for example, an index indicating the likelihood of converting the sentences of the inclusion criteria into structured data.

[0047] In the example of the display screen in Figure 9, the criteria items included in the exclusion criteria are associated with the conditions for each item. In addition, each item and its conditions are further associated with a score. The score is, for example, an index that indicates the likelihood of converting the sentences in the exclusion criteria into structured data.

[0048] The storage unit 16 stores, for example, data related to the process of generating reference indications. The storage unit 16 also stores, for example, dictionary data. The storage unit 16 also stores, for example, labeled models. The storage unit 16 also stores, for example, structured models. The storage unit 16 also stores, for example, generative models. The labeled models, structured models, and generative models may be stored in a storage means external to the selection support device 10.

[0049] The terminal device 20 is, for example, a terminal device used by a person who selects candidate clinical trial patients. The terminal device 20 acquires, for example, selection criteria for clinical trial patients. The selection criteria for clinical trial patients are input into the terminal device 20 by, for example, an operation by the person who selects candidate clinical trial patients. The terminal device 20 then outputs the selection criteria for clinical trial patients to the acquisition unit 11 of the selection support device 10.

[0050] The terminal device 20 acquires a reference instruction from, for example, the output unit 15 of the selection support device 10. Then, the terminal device 20 outputs the reference instruction to a display device (not shown). Furthermore, the terminal device 20 acquires structured selection criteria from, for example, the output unit 15 of the selection support device 10. Then, the terminal device 20 outputs the structured selection criteria to a display device (not shown).

[0051] The operation in the process of generating a reference designation will be described below with reference to Fig. 10, which is a diagram showing an example of the operation flow in the process of generating a reference designation.

[0052] The acquisition unit 11 acquires a sentence indicating the selection criteria for clinical trial patients (Step S11). The acquisition unit 11 acquires the sentence indicating the selection criteria from the terminal device 20, for example.

[0053] When the sentence indicating the selection criteria is acquired, the conversion unit 12 converts the sentence indicating the selection criteria into structured selection criteria (step S12).

[0054] When all sentences included in the selection criteria have been converted into structured selection criteria (Yes in step S13), the generation unit 14 generates reference instructions for information to be referenced when selecting clinical trial patients based on the structured selection criteria (step S14).

[0055] When the reference indication is generated, the output unit 15 outputs the generated reference indication (step S15). The output unit 15 outputs the generated reference indication to, for example, the terminal device 20. The output unit 15 may further output structured selection criteria. Furthermore, when the structured selection criteria do not include an item for which a reference indication is to be generated, the output unit 15 may output the structured selection criteria.

[0056] If there is a sentence that has not been converted in step S13 (No in step S13), the process returns to step S12, and the conversion unit 12 converts the sentence indicating the selection criteria that has not been converted into a structured selection criteria.

[0057] The selection support device 10 converts the selection criteria for clinical trial patients into structured selection criteria. Based on the structured selection criteria, the selection support device 10 generates reference instructions for information to be referenced in selecting clinical trial patients. The selection support device 10 then outputs the generated reference instructions. By generating reference instructions for information to be referenced in selecting clinical trial patients, for example, a person selecting clinical trial patients can confirm the information to be referenced in selecting clinical trial patients based on the reference instructions. Therefore, use of the selection support device 10 makes it easy to identify the information to be referenced in selecting clinical trial patients.

[0058] By identifying items for which reference instructions are generated based on the likelihood in the process of converting into structured selection criteria, for example, a person extracting clinical trial patient candidates can check detailed treatment information for items with low likelihood. Therefore, by identifying items for which reference instructions are generated based on the likelihood in the process of converting into structured selection criteria and generating the reference instructions, the accuracy of extracting patients who meet the selection criteria can be improved. Furthermore, by identifying items for which reference instructions are generated based on criteria set for each disease name or condition, the selection support device 10 can generate appropriate reference instructions according to the disease name or condition. Therefore, the selection support device 10 can further improve the accuracy of extracting patients who meet the selection criteria.

[0059] Furthermore, by generating the reference instructions in the form of a prompt used for input to a large-scale language model, it may be possible, for example, to extract a portion corresponding to the reference instruction from an electronic medical record using the large-scale language model. By extracting a portion corresponding to the reference instruction from an electronic medical record using the large-scale language model, a person selecting candidate clinical trial patients can easily refer to information necessary to determine, for example, whether a patient meets the selection criteria for the clinical trial. Therefore, the selection support device 10 can support the efficient extraction of patients who meet the selection criteria.

[0060] Furthermore, by generating reference instructions based on structured selection criteria, the selection support device 10 can generate reference instructions optimized according to the selection criteria for clinical trial patients. Furthermore, by making it easier to understand the information that should be referenced in patient selection, the person who selects patients who meet the selection criteria for the clinical trial can make appropriate decisions in selecting clinical trial patients.

[0061] Furthermore, by outputting the structured selection criteria together with a score indicating the likelihood of converting the text indicating the selection criteria into structured data, the person selecting candidate patients for clinical trials can check information about treatment by referring to the accuracy of the structured selection criteria. Therefore, by outputting the structured selection criteria together with the score, the person selecting candidate patients for clinical trials can appropriately determine the information that needs to be referred to.

[0062] The processes in the selection support device 10 may be distributed and executed among multiple information processing devices connected via a network. For example, the process of converting a sentence indicating the selection criteria into structured selection criteria and the process of generating reference indications may be executed in different information processing devices. It can be set as appropriate which information processing device executes each process in the selection support device 10.

[0063] Each process in the selection support device 10 can be realized by executing a computer program on a computer. Fig. 11 shows an example of the configuration of a computer 100 that executes a computer program that performs each process in the selection support device 10. The computer 100 includes a CPU (Central Processing Unit) 101, a memory 102, a storage device 103, an input / output I / F (Interface) 104, and a communication I / F 105.

[0064] The CPU 101 reads and executes computer programs for performing each process from the storage device 103. The CPU 101 may be configured with a combination of multiple CPUs. The CPU 101 may also be configured with a combination of a CPU and another type of processor. For example, the CPU 101 may be configured with a combination of a CPU and a graphics processing unit (GPU). The memory 102 is configured with a dynamic random access memory (DRAM) or the like, and temporarily stores computer programs executed by the CPU 101 and data being processed. The storage device 103 stores computer programs executed by the CPU 101. The storage device 103 is configured with, for example, a non-volatile semiconductor storage device. Other storage devices such as a hard disk drive may also be used for the storage device 103. The input / output I / F 104 is an interface that receives input from an operator and outputs display data, etc. The communication I / F 105 is an interface that transmits and receives data between the terminal device 20 and other information processing devices. The terminal device 20 may also have a configuration similar to that of the computer 100.

[0065] The computer program used to execute each process can also be stored and distributed on a computer-readable recording medium that non-temporarily stores data. Examples of recording media that can be used include magnetic tapes for recording data and magnetic disks such as hard disks. Optical disks such as CD-ROMs (Compact Disc Read Only Memory) can also be used as recording media. Non-volatile semiconductor storage devices can also be used as recording media.

[0066] Some or all of the above-described embodiments can be described as, but are not limited to, the following supplementary notes.

[0067] [Supplementary Note 1] A selection support device comprising: an acquisition means for acquiring text indicating selection criteria for clinical trial patients; a conversion means for converting the text indicating the selection criteria into structured selection criteria; a generation means for generating reference instructions indicating information to be referenced in selecting the clinical trial patients based on the structured selection criteria; and an output means for outputting the generated reference instructions.

[0068] [Supplementary Note 2] The selection support device according to Supplementary Note 1, wherein the structured selection criteria are data in which items indicating the content of each criterion included in a sentence indicating the selection criteria are associated with conditions for each of the items.

[0069] [Supplementary Note 3] The selection support device described in Supplementary Note 1 or 2, wherein the conversion means adds labels to terms related to the selection of the clinical trial patients included in the sentence indicating the selection criteria, and converts the sentence indicating the selection criteria into the structured selection criteria based on the added labels.

[0070] [Supplementary Note 4] The selection support device according to Supplementary Note 3, wherein the conversion means adds the labels based on a chronological relationship between the terms.

[0071] [Supplementary Note 5] The selection support device according to any one of Supplementary Notes 1 to 4, wherein the reference instruction includes a location where the information to be referenced is described and the content to be referenced.

[0072] [Supplementary Note 6] The selection support device according to Supplementary Note 5, wherein the reference instruction includes a location in an electronic medical record where the information to be referenced is written and the content to be referenced.

[0073] [Supplementary Note 7] The selection support device according to any one of Supplementary Notes 1 to 6, wherein the text indicating the selection criteria includes at least one of inclusion conditions and exclusion conditions for the clinical trial.

[0074] [Supplementary Note 8] The selection assistance device according to any one of Supplementary Notes 1 to 7, wherein the reference indication is a prompt used as input to a large-scale language model.

[0075] [Supplementary Note 9] The selection support device according to any one of Supplementary Notes 1 to 8, wherein the output means further outputs the structured selection criteria.

[0076] [Supplementary Note 10] The selection support device according to Supplementary Note 9, wherein the output means outputs a score indicating the likelihood of conversion into the structured selection criteria in association with each item of the structured selection criteria.

[0077] [Supplementary Note 11] The selection support device according to Supplementary Note 9 or 10, wherein the acquisition means acquires an item that is the target of the reference indication from among the selected selection criteria, and the generation means generates the reference indication for the item that is the target of the acquired reference indication.

[0078] [Supplementary Note 12] The selection support device according to Supplementary Note 3, wherein the conversion means adds labels to terms included in the selection criteria using a labeling model generated by machine learning.

[0079] [Supplementary Note 13] A selection support method comprising: acquiring a text indicating selection criteria for clinical trial patients; converting the text indicating the selection criteria into structured selection criteria; generating reference instructions indicating information to be referenced in selecting the clinical trial patients based on the structured selection criteria; and outputting the generated reference instructions.

[0080] [Supplementary Note 14] A recording medium that non-temporarily records a selection support program that causes a computer to execute the following processes: a process of acquiring text indicating selection criteria for clinical trial patients; a process of converting the text indicating the selection criteria into structured selection criteria; a process of generating reference instructions indicating information to be referenced in selecting the clinical trial patients based on the structured selection criteria; and a process of outputting the generated reference instructions.

[0081] Furthermore, some or all of the configurations described in Supplementary Notes 2 to 12, which are dependent on Supplementary Note 1, may also be dependent on Supplementary Notes 13 and 14 in the same dependent relationship as Supplementary Notes 2 to 12. Furthermore, not limited to Supplementary Notes 1, 13, and 14, some or all of the configurations described as Supplements may be made dependent on various hardware, software, various recording means for recording software, or systems, within the scope of each of the above-mentioned embodiments.

[0082] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above-described embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure. Furthermore, each embodiment can be combined with other embodiments as appropriate.

[0083] REFERENCE SIGNS LIST 10 Selection support device 11 Acquisition unit 12 Conversion unit 13 Identification unit 14 Generation unit 15 Output unit 16 Storage unit 20 Terminal device 100 Computer 101 CPU 102 Memory 103 Storage device 104 Input / output I / F 105 Communication I / F

Claims

1. An acquisition means for acquiring a text indicating the selection criteria for clinical trial patients, a conversion means for converting the text indicating the selection criteria into structured selection criteria, a generation means for generating a reference instruction indicating information to be referred to in the selection of the clinical trial patients based on the structured selection criteria, and an output means for outputting the generated reference instruction. A selection support device comprising:

2. The selection support device according to claim 1, wherein the structured selection criteria are data associating items indicating the content of each criterion included in the text indicating the selection criteria and the conditions for each of the items.

3. The conversion means according to claim 1 or 2, which adds a label to terms related to the selection of the clinical trial patients included in the text indicating the selection criteria, and converts the text indicating the selection criteria into the structured selection criteria based on the added label.

4. The selection support device according to claim 3, wherein the conversion means adds the label based on the temporal relationship of the terms.

5. The selection support device according to any one of claims 1 to 4, wherein the reference instruction includes a location where the information to be referred to is described and the content to be referred to.

6. The selection support device according to claim 5, wherein the reference instruction includes a location where the information to be referred to is described and the content to be referred to in an electronic medical record.

7. The selection support device according to any one of claims 1 to 6, wherein the text indicating the selection criteria includes at least one of inclusion criteria and exclusion criteria for a clinical trial.

8. The selection support device according to any one of claims 1 to 7, wherein the reference instruction is a prompt used as an input to a large language model.

9. The selection support device according to any one of claims 1 to 8, wherein the output means further outputs the structured selection criteria.

10. The selection support device according to claim 9, wherein the output means outputs in association with each item of the structured selection criteria a score indicating the likelihood of conversion to the structured selection criteria.

11. The acquisition means acquires an item that is the target of the reference instruction among the selected selection criteria, and the generation means generates the reference instruction for the item that is the target of the acquired reference instruction. The selection support device according to claim 9 or 10.

12. The selection support device according to claim 3, wherein the conversion means adds a label to a term included in the selection criteria using a labeled model generated by machine learning.

13. A selection support method, comprising: obtaining a sentence indicating selection criteria for a clinical trial patient; converting the sentence indicating the selection criteria into structured selection criteria; generating a reference instruction indicating information to be referred to in the selection of the clinical trial patient based on the structured selection criteria; and outputting the generated reference instruction.

14. The selection support method according to claim 13, wherein the structured selection criteria are data associating items indicating the content of each criterion included in the sentence indicating the selection criteria with the conditions of each item.

15. The selection support method according to claim 13 or 14, comprising: adding a label to a term related to the selection of the clinical trial patient included in the sentence indicating the selection criteria; and converting the sentence indicating the selection criteria into the structured selection criteria based on the added label.

16. The selection support method according to claim 15, wherein the label is added based on a temporal relationship of the terms.

17. The selection support method according to any one of claims 13 to 16, wherein the reference instruction includes a location where the information to be referred to is described and the content to be referred to.

18. The selection support method according to claim 17, wherein the reference instruction includes a location where the information to be referred to is described and the content to be referred to in an electronic medical record.

19. The selection support method according to any one of claims 13 to 18, wherein the sentence indicating the selection criteria includes at least one of inclusion criteria and exclusion criteria for a clinical trial.

20. A recording medium non-temporarily recording a selection support program for causing a computer to execute: a process of obtaining a sentence indicating selection criteria for a clinical trial patient; a process of converting the sentence indicating the selection criteria into structured selection criteria; a process of generating a reference instruction indicating information to be referred to in the selection of the clinical trial patient based on the structured selection criteria; and a process of outputting the generated reference instruction.

Images