Side effect information searching system, side effect information searching device, side effect information searching method, and computer program

Abstract

To enable a fuzzy search with various ways of writing on the basis of terms indicating side effects for searching in conducting a search related to side effects of medicine.SOLUTION: A server 10 specifies, for side effect terms for searching, synonymous side effect terms from a synonymous side effect table 14, for each of the side effect terms for searching and the synonymous side effect terms, decomposes the term into a plurality of words on the basis of a side effect word table 15, specifies a relationship in which the words obtained through the decomposition are synonyms for each other, from a synonymous word table 16, couples the words in the relationship that they are synonyms for each other to create side effect terms written in a different form, so that they are written in a different form from the side effect terms for searching and the synonymous side effect terms, and conducts a search with the side effect terms for searching, synonymous side effect terms, and side effect terms written in the different form.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a side effect information search system, a side effect information search device, a side effect information search method, and a computer program that enable fuzzy searches using various notations for terms that represent side effects and symptoms used as search words when searching a data group that includes words that indicate side effects and symptoms of pharmaceuticals, thereby enabling flexible detection of pharmaceutical-related literature information and patient information. [Background technology]

[0002] Numerous studies have been conducted on the side effects and symptoms (including early symptoms) of pharmaceuticals, and the results and outcomes of these studies are published in medical literature (including papers, guidelines, books, and drug package inserts). To make such literature widely available to researchers and medical professionals, there are increasing examples of data being made accessible by being digitized and made searchable, and included in big data. In addition, patient records such as medical records and medication histories are also being digitized and made searchable in the systems of various medical institutions. To make effective use of such digitized data, various technologies and mechanisms are being developed for various processes such as searching.

[0003] For example, in Patent Document 1 below, in a screening operation in which information is extracted from documents such as medical papers to determine whether it corresponds to a side effect of a drug, drug information that identifies the drug, symptom information related to the symptoms, and change information that indicates a change in the state are extracted for each sentence from the documents such as papers. In this process, a side effect dictionary such as MedDRA (Medical Dictionary for Regulatory Activities), a specialized medical dictionary such as a medical idiom dictionary, a synonym dictionary, etc. are used to generate structured information that indicates the drug name, symptoms, and changes for the documents such as medical papers, and changes in wording that absorb variations in the spelling of symptoms, etc. (See the description in paragraphs 0040-0071, etc., of Patent Document 1).

[0004] Furthermore, Patent Document 2 below discloses the use of a drug information DB (database) including drug names and side effect disease names, a side effect disease name DB for standardizing the various expressions of side effect disease names into a predetermined expression, and a medication history information DB for storing information on drugs prescribed to patients, in order to identify drugs that cause side effects (see paragraphs 0024-0033, Figure 2, etc. of Patent Document 2).

[0005] Furthermore, Patent Document 3 listed below discloses the creation of a search formula based on systematically accumulated information on example expressions related to side effects. Specifically, it uses a database (side effect expression DB) in which side effect expressions (expressions that express the symptoms of side effects as experienced by ordinary people, side effect types), state expressions, etc. are associated with drug IDs (see Figures 4, 14, etc. of Patent Document 3), and discloses a process of breaking down sentences into various parts of speech as a result of morphological analysis and extracting candidate drug names, side effect expressions, etc. (see Figures 7, 12, etc. of Patent Document 3), and furthermore, when a drug name is entered, a list of summary information for each side effect type can be displayed (see Figure 26 of Patent Document 3).

[0006] Furthermore, Patent Document 4 below discloses that, in order to determine which drug or combination of drugs is causing the side effects based on the initial symptoms, side effect information for the corresponding drug information is read out from a side effect information database based on the initial symptom information and the corresponding drug information (see paragraphs 0021-0029, etc., of Patent Document 4).

[0007] Furthermore, Patent Document 5 listed below discloses the use of a pharmaceutical database that associates drug information including pharmaceutical names with side effect information, and an early symptom database that associates drug information including pharmaceutical names with side effect information and early symptom information, in order to search for and output drug information and early symptom information based on side effect information (see Figures 10, 11, etc. of Patent Document 5). [Prior art documents] [Patent documents]

[0008] [Patent Document 1] Patent Publication No. 2021-2309 [Patent Document 2] Japanese Patent Application Publication No. 2018-147463 [Patent Document 3] Japanese Patent Application Laid-Open No. 2008-234003 [Patent Document 4] Japanese Patent Application Laid-Open No. 2006-107321 [Patent Document 5] Japanese Patent Application Laid-Open No. 2004-290377 Summary of the Invention [Problem to be solved by the invention]

[0009] The technical contents disclosed in the above-mentioned Patent Documents 1 to 5 are useful for screening medical documents for searching for drugs corresponding to side effects or symptoms using search terms that represent drug side effects or symptoms.Similarly, the technical contents disclosed in the above-mentioned Patent Documents 1 to 5 are useful for screening medical documents for searching for drug side effects or symptoms using search terms that represent drug side effects or symptoms using search terms that represent drug side effects or symptoms.

[0010] However, some terms that represent side effects or symptoms are created by combining multiple words. In the case of such terms that combine multiple words, there are synonyms and different spellings even for the combined words (words). Therefore, when these words are combined to create terms that represent side effects or symptoms with the same meaning, there will be many different spelling patterns for the terms that represent the same side effects or symptoms, which creates a problem in that the technical content disclosed in Patent Documents 1-5 cannot address all of these issues.

[0011] For example, the term "leukopenia," which indicates a side effect, is a combination of two words, "white blood cell" and "decrease." Another example of how to write "leukocyte" is "WBC (White Blood Cell)," and an example of a synonym for "decrease" is "decrease." Therefore, by appropriately combining these words, three different terms can be created that have the same meaning as "leukopenia": "decrease in white blood cells," "decrease in WBC," and "decrease in WBC." It is entirely possible that these three terms, like "leukopenia," will be included in the data set to be searched.

[0012] On the other hand, the technology of Patent Document 1 only absorbs variations in the spelling of a single word, such as "edema" for "swelling" by referring to a specialized dictionary or a thesaurus (see paragraphs 0043 and 0054 of Patent Document 1). Therefore, since the multiple words that make up the above-mentioned "leukopenia" or terms that are spelled in various combinations of words with equivalent meanings are not included in specialized dictionaries or thesauruses, it is not possible to use terms with other spelling patterns, such as "leukocyte decrease," "WBC decrease," or "WBC decrease," as search words, which raises concerns about missed searches.

[0013] Similarly, in the technology of Patent Document 2, a database of side effect disease names is used to standardize the various expressions of side effect disease names into a predetermined expression (see the description in paragraph 0030 of Patent Document 2, etc.).

[0014] However, because the specific contents of the database of adverse drug reaction disease names have not been disclosed, it is unclear how the multiple patterns of notation for terms that combine multiple words will be standardized into a predetermined expression, and concerns about missed searches cannot be dispelled.

[0015] Furthermore, Patent Document 3 explains that parts of speech such as verbs and adjectives vary in notation depending on the conjugation form, and that the notation converted into a set conjugation form is displayed as supplementary information (see paragraphs 0046, 0047, etc. of Patent Document 3), but this simply converts the conjugation form of verbs, adjectives, etc., and does not address the issue of identifying multiple notation patterns for a term that combines multiple words and using them as search words to perform search processing, etc. Patent Documents 4 and 5 do not disclose multiple notation patterns for a term that combines multiple words, and therefore cannot resolve the concern of missed searches that arise in search processing for terms that are expected to have multiple patterns.

[0016] Furthermore, in actual medical settings, when recording side effects and symptoms in medical records or medication histories, they are sometimes written in hiragana or katakana to indicate how to read them, rather than in kanji, and these hiragana and katakana entries remain even when the data is digitized.

[0017] However, when terms for side effects or symptoms that are normally written in kanji are written in hiragana or katakana, the specialized dictionaries and synonym dictionaries used in Patent Document 1 and the like may not include the hiragana or katakana spellings, which raises concerns that the search may miss such cases as well.

[0018] This also applies to the database of side effect disease names used in Patent Document 2. It is unclear whether the database of side effect disease names in Patent Document 2 also includes hiragana and katakana spellings for terms for side effects and symptoms written in kanji, so there remains concern that searches may be missed in this regard as well.

[0019] The present invention has been made in consideration of the above circumstances, and aims to provide a side effect information search system, a side effect information search device, a side effect information search method, and a computer program that, when searching for terms that represent side effects or symptoms that are combinations of multiple words, enable searches to be performed including other expressions with the same meaning, thereby enabling comprehensive ambiguous searches and avoiding search omissions.

[0020] Another object of the present invention is to provide a side effect information search system, a side effect information search device, a side effect information search method, and a computer program that enable flexible search and extraction of information describing side effects and symptoms to be searched for, even when the information includes information indicating medical literature or information about patients. [Means for solving the problem]

[0021] In the present invention, when a search is performed using a search side effect term, the search process for a data group is performed by referring to tables such as a synonymous side effect table, a side effect word table, and a synonymous word table (which correspond to information groups that show the correspondence between information related to words) based on such search side effect terms, and using the generated side effect terms as search words.In this case, the generated side effect terms include at least side effect terms that are spelled differently from the search side effect terms, and these side effect terms with different spellings are also used as search words, thereby making it possible to prevent search omissions that occurred in conventional technology.

[0022] In the present invention, two patterns are assumed for generating side effect terms to be used as search words based on side effect terms for searches. The first pattern is when there are no restrictions on the words that make up the side effect terms for searches. In this case, based on the synonymous side effect table, multiple word-combined side effect terms (side effect terms formed by combining multiple words) are identified as synonymous side effect terms for the side effect terms for searches. Various processes are then performed on the multiple word-combined side effect terms thus identified to generate side effect terms (including side effect terms with different spellings) to be used as search words.

[0023] In this first pattern, there are no restrictions on the words that make up the side effect terms used for searching (side effect terms used for searching may consist of only one word or may consist of multiple words), so side effect terms used for searching can be freely selected.

[0024] The second pattern according to the present invention is to use a side effect term for a search that is a combination of multiple words, and to identify a word combination side effect term (at least one or more word combination side effect terms) as a synonymous side effect term for such a side effect term for a search based on a synonymous side effect table.

[0025] Then, by performing various processes on the thus identified word combination side effect terms and the side effect terms for search, side effect terms (including side effect terms with different spellings) to be used as search words are generated.

[0026] In this second pattern, the side effect term used for the search is a combination of multiple words, but as long as there is at least one synonymous side effect term (word-combined side effect term) identified based on the synonymous side effect table, the search process can be performed without any problems. Therefore, even if the number of synonymous side effect terms (word-combined side effect terms) for the side effect term used for the search is limited, the search process can be performed. In addition to medical terms used in diagnostic names by doctors, side effect terms include terms indicating side effects used in MedDRA / J (ICH International Medical Dictionary / Japanese Edition), CTCAE (Common Terminology Criteria for Adverse Events), drug package inserts, interview forms, medical dictionaries, etc., and their synonyms.

[0027] In the present invention, groups of synonymous words are identified based on the synonym word table, and one word is extracted from each of the identified groups of synonymous words. When generating side effect terms by combining these words in the order of combination shown in the side effect word table, all possible combinations are considered, so that all side effect terms are generated in a brute force manner using each extracted word, thereby ensuring that no search misses anything.

[0028] In the present invention, in the side effect word table, each word used in a word combination side effect term is indicated with a level value corresponding to the importance in expressing the meaning of the word combination side effect term. When multiple side effect terms are generated, each side effect term is used as a search word in an order corresponding to the level value of the word contained in each side effect term. This makes it possible to perform search processing in an order close to the meaning of the word combination side effect term, thereby improving the efficiency of the search processing to obtain search results desired by the user.

[0029] In the present invention, a drug link table showing drug names corresponding to side effect terms is used, and drug names corresponding to side effect terms found in the search process are identified based on the drug link table and output processed.One of the drug names output from the output process is accepted, and information listing both the accepted drug name and the side effect terms for that drug name is extracted from the data group.Therefore, even if there are many drug names corresponding to side effect terms, the user can select to limit the drug names used to extract information from the data group, and as a result, the information extracted from the data group is narrowed down, thereby improving the efficiency of screening tasks, etc.

[0030] In particular, in the present invention, the data group to be searched includes literature information or patient information containing patient names, and any drug name is accepted from the drug names that have been output, and ultimately, literature information or patient information containing patient names that includes the side effect terms that were matched in the search process and information containing the accepted drug name is extracted from the data group.

[0031] Therefore, when extracting literature information containing the side effect terms that were matched in the search process and the accepted drug names from the literature information contained in the data group, it is possible to find literature information (papers, guidelines, books, drug package inserts, etc.) related to both the side effect terms used in the search and the drug names corresponding to those side effect terms from the large amount of literature information contained in the data group to be searched, thereby further improving the efficiency of screening work to find literature information related to terms that describe side effects and drugs from the large amount of literature information contained in the data group to be searched.

[0032] Furthermore, when extracting patient information containing the side effect terms matched in the search process and the names of the drugs received from the patient information contained in the data group, it becomes possible to efficiently find patient information (patient information according to electronic medical records, patient information according to electronic medical histories, etc.) that is related to both the side effect terms used in the search and the drug names corresponding to those side effect terms from the large amount of patient information contained in the data group to be searched.This makes it possible, for example, to quickly find patients who are likely to experience a specific side effect, and is useful for quickly responding to such patients.

[0033] In the present invention, side effect terms included in the drug link table are used as side effect terms for searches, so that the search process can be carried out using the side effect terms included in the drug link table as search words. As a result, the search process can be performed without the user having to input search words, and the search process can be automated based on the side effect terms in the drug link table. Note that the side effect terms included in the drug link table used as search words in this case may be all of the side effect terms included in the drug link table, or side effect terms of a certain level (for example, the most important level) may be used. When all of the side effect terms included in the drug link table are used, for example, if the drug link table contains 1,000 side effect terms, the above-mentioned processing will be performed for each of the 1,000 side effect terms.

[0034] In the present invention, side effect terms and drug names for the search are accepted, and drug names corresponding to the side effect terms found in the search process are identified based on a drug link table. From among the identified drug names, drug names that are the same as the accepted drug name are identified. This makes it possible to narrow down the number of drug names corresponding to the side effect terms found in the search process using the drug names entered by the user, thereby enabling efficient screening work to be carried out in accordance with the drug names confirmed by the user.

[0035] In the present invention, when a symptom term indicating a symptom is written in the information contained in the data group to be searched, and a search is performed based on a symptom term combining multiple words, the symptom term to be used as the search word is identified individually based on the symptom table, and then ``symptom words (words that best indicate the meaning of the symptom)'' and ``synonymous words (synonymous words)'' are identified from the symptom synonym table.Furthermore, for each of the ``symptom words'' and ``synonymous words,'' ``pronunciation words (words that indicate pronunciation)'' are identified based on the symptom word pronunciation table, and the data group is searched (word search process) using these ``symptom words,'' ``synonymous words,'' and ``pronunciation words.''Therefore, it is possible to search not only using terms indicating side effects, but also using terms indicating symptoms, and it is possible to search the data group using a wide range of terms, thereby reducing the risk of missing results.

[0036] In the present invention, symptom terms containing words found in the above-mentioned word search process are identified based on the symptom table, and side effect terms corresponding to the identified symptom terms are identified based on the side effect symptom link table. This makes it possible to find side effect terms corresponding to the symptom terms obtained through the word search process, and allows one to confirm both the symptoms found in the search results and the side effects related to those symptoms.

[0037] In the present invention, when the side effect symptom link table has two or more symptom terms corresponding to one side effect term and there are multiple symptom terms corresponding to the side effect term identified in the side effect symptom link table, the side effect symptom link table outputs the ratio of the number of symptom terms that use words found to be included in the search process to the number of symptom terms corresponding to the side effect term identified in the side effect symptom link table, so that the probability that a symptom term containing a word used in the word search process could be a symptom term corresponding to the side effect term can be grasped in an objective numerical value, that is, a ratio.

[0038] In this invention, if the symptom terms contained in the symptom table are used as the symptom terms for search, a search can be performed without the user having to input the symptom terms as search words, thereby automating the search based on the symptom terms in the symptom table. In this case, the symptom terms contained in the symptom table to be used as search words may be all of the symptom terms contained in the symptom table, as in the case where the side effect terms contained in the drug link table described above are used as side effect terms for search, or a certain level of symptom terms may be used.

[0039] In the present invention, since the symptom terms contained in the symptom table are used as the symptom terms for searching, the search process can be carried out using the symptom terms contained in the symptom table as search words. As a result, even in the search process related to symptom terms, automatic search processing can be performed without the user having to input search words. In addition, in the present invention, symptom terms and drug names for search are accepted, and drug names corresponding to the identified side effect terms are identified based on a drug link table, and from among the identified drug names, the drug names are limited (identified) to those that are the same as the accepted drug name.Therefore, even if there are many drug names corresponding to the side effect terms, the drug names are narrowed down by the drug name entered by the user, and subsequent search processing of data groups can be performed efficiently. [Effects of the Invention]

[0040] In the present invention, even if the side effect term used for searching is composed of one word or multiple words, the search process is performed using multiple side effect terms with the same meaning but different spellings, using a synonymous side effect table, a side effect word table, and a synonymous word table, thereby reducing search omissions compared to conventional technology. Furthermore, in the present invention, side effect terms are generated by combining each word extracted from each identified synonymous word group in a round-robin manner, and then search processing is performed, thereby reliably preventing missed searches.

[0041] Furthermore, in the present invention, when multiple side effect terms are generated, each side effect term is used as a search word in an order according to the level value of the words contained in each side effect term, so that search processing can be performed in an order close to the meaning of the terms used in the search, thereby realizing efficient search processing.

[0042] In the present invention, drug names corresponding to side effect terms found in the search process are identified and output based on a drug link table, allowing the user to check how many drug names exist that correspond to the side effect terms.In addition, the output drug names can be selected by the user, and information containing both the accepted drug name and the side effect terms for that drug name is extracted from the data group.This makes it possible to detect side effect terms and information related to those side effect terms and corresponding to the drug names specified by the user from a large amount of information, contributing to the efficiency of screening work.

[0043] In particular, in the present invention, literature information or patient information that includes both the side effect terms that were matched in the search process and the drug names selected by the user is extracted from the literature information contained in the data group.Therefore, when extracting literature information, it is possible to easily find literature information that is related to the side effect terms and includes content related to the drug names narrowed down by the user.Furthermore, when extracting patient information, it is possible to efficiently detect patients that are related to the side effect terms and related to the drug names narrowed down by the user, and this can be used as a kind of alert to identify patients whose situation requires attention.

[0044] Furthermore, in the present invention, the search process is performed by using the side effect terms contained in the drug link table as side effect terms for searching, so the search process can be performed without the user having to input search words, and the search process can be automated based on the side effect terms in the drug link table.

[0045] In addition, in the present invention, side effect terms and drug names for search are accepted, a search process is performed based on such search terms, and a drug name that is the same as the accepted drug name is identified from among the drug names corresponding to the side effect terms that are hit in the search process.Therefore, even if there are many drug names corresponding to the side effect terms that are hit, it is possible to narrow down the search to the drug name set by the user.

[0046] In the present invention, when information containing symptom terms indicating symptoms is included in the data set to be searched, searches can be performed using "symptom words," "synonymous words," and "pronunciation words" as search terms, thereby enabling searches to be performed that are consistent with the notation of symptoms recorded in medical records, medication histories, etc. in actual medical settings. Furthermore, side effect terms corresponding to the symptom terms obtained through the search process are identified based on the side effect symptom link table, allowing both the symptoms resulting from the search and the side effects associated with such symptoms to be confirmed. Furthermore, side effect terms corresponding to the symptom terms obtained through the search process are used as the side effect terms for search, allowing for smooth linkage from processing based on symptom terms to processing based on side effect terms, thereby helping to effectively perform screening tasks based on symptom terms and side effect terms. In addition, in the present invention, by using symptom terms included in the symptom table as symptom terms for search, search processing based on symptom terms can be performed without inputting search words, and search processing for symptom terms can also be automated.

[0047] Furthermore, in the present invention, when the side effect symptom link table contains two or more symptom terms corresponding to one side effect term and there are multiple symptom terms corresponding to the side effect term identified in the side effect symptom link table, the number of identified symptom terms that use words found to be included in the data group search process and the ratio of the number of symptom terms corresponding to the side effect term identified in the side effect symptom link table are output.Therefore, for symptom terms containing words used in the word search process, the probability that they could be symptom terms corresponding to the side effect term can be presented as an objective numerical value in the form of a ratio, which can be used to understand the accuracy of search results based on symptom terms.

[0048] Furthermore, in the present invention, the symptom terms contained in the symptom table are used as the symptom terms for searching, so that the search process can be performed without the user having to input search words, and the search process can be automated based on the symptom terms in the symptom table. Furthermore, the present invention accepts symptom terms and drug names for search, identifies side effect terms from the symptom terms obtained as a result of search processing based on the accepted symptom terms, and then narrows down multiple drug names corresponding to the identified side effect terms using the accepted drug names.Therefore, even if there are many drug names corresponding to the side effect terms, the search can be narrowed down using the drug names entered by the user, making it easier to obtain results that meet the user's needs. [Brief explanation of the drawings]

[0049] [Figure 1] 1 is a schematic diagram showing an example of the overall system configuration including a side effect information search system according to a first embodiment of the present invention. [Figure 2] (a)-(c) show examples of data groups to be searched, where (a) is a schematic diagram showing part of the big data, (b) is a schematic diagram showing part of X Hospital's electronic medical record database, and (c) is a schematic diagram showing part of Y Pharmacy's electronic medical history database. [Figure 3] FIG. 2 is a block diagram showing the main internal configuration of the server device. [Figure 4] 10 is a diagram showing an example of partial contents of a drug link table. [Figure 5]10 is a diagram illustrating an example of partial contents of a synonymous side effect table. [Figure 6] 10 is a diagram showing an example of partial contents of a side effect word table. [Figure 7] 10 is a diagram showing an example of partial contents of a synonymous word table. [Figure 8] (a) is a schematic diagram showing the co-occurrence of each word, (b) is a schematic diagram showing the synonymous relationship of each word, and (c) is a schematic diagram showing the generation status of side effect terms with different notations. [Figure 9] FIG. 2 is a block diagram showing the main internal configuration of the terminal device. [Figure 10] 1A is a schematic diagram showing a login screen, and FIG. 1B is a schematic diagram showing a search target confirmation screen. [Figure 11] FIG. 10 is a schematic diagram showing a search word setting screen. [Figure 12] 1A is a schematic diagram showing a search type setting screen, and FIG. 1B is a schematic diagram showing an automatic search confirmation screen. [Figure 13] FIG. 10 is a schematic diagram showing a pharmaceutical product name list screen. [Figure 14] FIG. 2 is a schematic diagram showing a document search result screen according to the first embodiment. [Figure 15] FIG. 2 is a schematic diagram showing an electronic medical record search result screen according to the first embodiment. [Figure 16] FIG. 3 is a schematic diagram showing an electronic medical history search result screen according to the first embodiment. [Figure 17] 1 is a first flowchart showing the processing steps of the side effect information search method according to the first embodiment. [Figure 18] 2 is a second flowchart showing the processing procedure of the side effect information search method according to the first embodiment. [Figure 19] 3 is a third flowchart showing the processing procedure of the side effect information search method according to the first embodiment. [Figure 20] 10 is a diagram showing an example of partial contents of a synonymous side effect table according to a modified example. [Figure 21]10 is a diagram showing an example of partial contents of a merge table that combines the contents of the synonymous side effect table and the contents of the side effect word table. [Figure 22] FIG. 1 is a schematic diagram showing the overall system configuration including a side effect information search system according to a second embodiment of the present invention. [Figure 23] 10 is a diagram showing an example of partial contents of a side effect symptom link table. [Figure 24] 10 is a diagram illustrating an example of partial contents of a symptom table. [Figure 25] 10 is a diagram showing an example of partial contents of a symptom synonym word table. [Figure 26] 10 is a diagram showing an example of partial contents of a symptom word pronunciation table. [Figure 27] FIG. 11 is a schematic diagram showing a search word setting screen according to the second embodiment. [Figure 28] FIG. 11 is a schematic diagram showing a document search result screen according to the second embodiment. [Figure 29] FIG. 11 is a schematic diagram showing an electronic medical record search result screen according to the second embodiment. [Figure 30] FIG. 11 is a schematic diagram showing an electronic medical history search result screen according to the second embodiment. [Figure 31] 4 is a fourth flowchart showing the processing procedure of the side effect information search method according to the second embodiment. [Figure 32] 5 is a fifth flowchart showing the processing procedure of the side effect information search method according to the second embodiment. [Figure 33] 6 is a sixth flowchart showing the processing procedure of the side effect information search method according to the second embodiment. [Figure 34] FIG. 11 is a schematic diagram showing a search word setting screen according to a modified example of the second embodiment. [Figure 35] FIG. 10 is a schematic diagram showing the overall system configuration including the side effect information search system according to the third embodiment of the present invention. [Figure 36] 10 is a diagram showing an example of partial contents of a common name link table. [Figure 37]FIG. 10 is a schematic diagram showing the overall system configuration including the side effect information search system according to the fourth embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION [Example]

[0050] 1 is a schematic diagram showing an overall system configuration as an example of an information retrieval system 1 according to a first embodiment (Example 1) of the present invention. The information retrieval system 1 is composed of a server device 10 that provides search results and the like, and a user terminal device 30. The server device 10 and the terminal device 30 work together to perform various processes such as search processing using a search word, and in this embodiment, a group of data (e.g., big data 5) stored in a cloud (e.g., a cloud system 4) or the like is the target of the search processing using the search word.

[0051] The data group to be searched includes information describing names of pharmaceuticals (drug names) and side effect terms indicating the side effects of the pharmaceuticals, and information describing side effect terms found in the search and names of pharmaceuticals corresponding to those side effect terms (e.g., literature information, patient information, etc.) is finally presented to the user. Note that in this embodiment, user registration is required to receive services from the information retrieval system 1, and the user registers by sending user information such as name, telephone number, email address, and date of birth to the server device 10 of the information retrieval system 1, and a user ID for identifying the user is issued.

[0052] The main processing according to the present invention is executed by a server device 10 (corresponding to a side effect information retrieval device), and each of the tables 13-16, etc. used to perform the specific processing according to the present invention is stored in advance in a table DB (database system) 11. A combination of such table DB 11 and server device 10 functions as a side effect information retrieval system 9 according to the present invention. Note that while FIG. 1 shows a configuration example of the side effect information retrieval system 9 in which the table DB 11 is incorporated into the server device 10, it is of course also possible to construct a side effect information retrieval system 9 in which the server device 10 and the table DB 11 are separate devices (hardware configurations) and the two devices are connected via a network, so that the server device 10 can use each of the tables 13, etc. stored in the table DB 11.

[0053] Furthermore, the server device 10 and the terminal device 30 are capable of communicating with each other via a network NW, and the server device 10 is also capable of accessing a group of data to be searched (for example, a cloud system 4 related to big data 5) via the communication network. Note that in order to provide an overview of the information retrieval system 1, only one terminal device 30 on the user side is shown in Fig. 1, but in reality, many terminal devices 30 of many users are connected to the network NW and are able to access the server device 10, and each of these users is able to receive the information retrieval service according to the present invention.

[0054] The cloud is constructed, for example, by a cloud system 4 existing on a network, and such cloud system 4 can be accessed via a cloud-related website, etc. Note that, since the configuration of a cloud is well known, a description of the cloud system 4 will be omitted. Furthermore, the cloud system 4 (and big data 5) is not included in the configuration of the information retrieval system 1 or the side effect information retrieval system 9, but is simply a search target (processing target) of the information retrieval system 1 or the side effect information retrieval system 9.

[0055] In addition to the cloud big data 5, other data groups that can be accessed by the server device 10 can also be set as search targets. In this embodiment, the user can select what can be such targets (files stored in a storage device connected to the network can also be selected as processing targets).

[0056] Specific examples of the information contained in data groups that can be searched include medical literature information (various medical papers, books, book reviews, magazines, drug package inserts, etc.), patient information with patient names (electronic medical records, electronic medication histories, etc.), etc. Medical literature information can be accessed through literature sites, etc., while patient information with patient names can be accessed through sites related to database systems that correspond to data groups that contain such information.

[0057] Possible users of the information search service according to this embodiment include medical professionals (doctors, nurses, pharmacists, etc.), patients themselves, medical researchers, etc. Furthermore, possible use cases of the information search service according to this embodiment include screening work to check for drugs according to side effects, narrowing down patients according to side effects, outputting alerts, and supporting research on side effects, etc.

[0058] FIG. 2(a) shows an overview of a data group corresponding to big data 5 stored in a cloud system 4, which is an example of a search target in the present invention. This big data 5 is a collection (data group) containing a large amount of literature information as data. In this embodiment, it includes information related to a large number of medical-related literature (medical-related literatures A, B, C, etc.). As described above, these medical-related literatures A, B, C, etc. include various medical-related papers, books, book reviews, magazines, drug package inserts, etc., and include information such as the name of the drug, side effect terms indicating the side effects of the drug, and symptom terms indicating symptoms (including initial symptoms). In addition to the content of the literature, medical-related literatures A, B, C, etc. also include, as appropriate, an outline of the content, the publication date, the author, the publisher, etc.

[0059] 2(b) shows an outline example of another type of data group that can be searched in this embodiment, and shows an X Hospital electronic medical record DB (database) 6 managed by X Hospital (X Hospital's electronic medical record system). This X Hospital electronic medical record DB 6 contains information indicating the contents of each patient's electronic medical record (patient information including patient name, etc.), such as the electronic medical record of patient K1, the electronic medical record of patient K2, and the electronic medical record of patient K3, and each of these patient electronic medical records includes patient identification information (patient name, patient telephone number, patient identification number, etc.), the name of the attending physician, information regarding the patient's treatment, etc. (including the patient's symptoms, etc.), information regarding the medicines administered to the patient (information regarding the medicine's name and terms related to side effects, etc., of the medicine indicated by the medicine's name), etc.

[0060] 2(c) shows an example of a summary of yet another type of data group that can be searched in this embodiment, showing the Y Pharmacy electronic medical history DB7 managed by Y Pharmacy (Y Pharmacy's electronic medical history system). This Y Pharmacy electronic medical history DB7 contains information indicating the contents of each patient's electronic medical history (patient information including patient name, etc.), such as the electronic medical history of patient C1, the electronic medical history of patient C2, and the electronic medical history of patient C3, and each of these patient electronic medical histories includes patient identification information (patient name, patient telephone number, patient identification number, etc.), historical information regarding the medicines prescribed to the patient (information regarding the medicine names and terms related to side effects of the medicines indicated by the medicine names), etc.

[0061] FIG. 3 is a block diagram showing the main configuration of the server device 10 in the information retrieval system 1. The server device 10 of this embodiment is a general-purpose server computer, and corresponds to a processing device (side effect information retrieval device) that performs the main processing of the present invention. Note that it is also possible to combine multiple server devices to construct a processing device (server device) according to this embodiment by performing distributed processing or the like for the processing according to the present invention (even when multiple devices are combined in this way, the entire configuration of the combined devices corresponds to the processing device (server device as a side effect information retrieval device) according to the present invention). Furthermore, as described above, the server device 10 of this embodiment incorporates a table DB 11 including various tables 13, etc., and the server device 10 with the table DB 11 incorporated functions as the side effect information retrieval system 9.

[0062] The server device 10 serves as the side effect information search device of the present invention, and is configured such that various devices are connected to an MPU 10a, which performs overall control and various processing, via internal connection lines 10h. The various devices include a communication module 10b, a RAM 10c, a ROM 10d, an input interface 10e, an output interface 10f, and a memory unit 10g.

[0063] The communication module 10b is a communication device equivalent to a connection module with the network NW, and conforms to a required communication standard (for example, a LAN module). The communication module 10b is connected to the network NW via a required communication device (not shown, such as a router) and enables communication with various communication devices such as the cloud system 4 and the terminal device 30. The RAM 10c temporarily stores contents, files, etc. associated with the processing of the MPU 10a, and the ROM 10d stores programs, etc., that define the basic processing contents of the MPU 10a.

[0064] The input interface 10e is connected to a keyboard, mouse, etc. for receiving operation instructions from a system administrator of the server device 10, and performs processing to receive operation instructions from the system administrator and transmit them to the MPU 10a. The output interface 10f is connected to a display (display output device), and outputs the contents of the processing performed by the MPU 10a to the display, allowing the system administrator to check the current processing contents, etc.

[0065] The storage unit 10g stores programs, databases, etc. In this embodiment, it stores a server OS program P and a search program 12 as programs, and a table DB 11, a user database 17, a screen database 18, etc. as databases. The table DB 11 also includes a drug link table 13, a synonymous side effect table 14, a side effect word table 15, and a synonymous word table 16. The server device 10 (MPU 10a) accesses each of the tables 13-16 and each of the databases 17 and 18 included in the table DB 11 stored in the storage unit 10g according to the processing status, and performs various processes based on the information stored in each table and database.

[0066] The server OS program P is a program corresponding to the operating system of the server device 10, and defines various processes to be performed by the server computer. The search program 12 is a computer program defining various processes according to the present invention, and the MPU 10a executes the processes defined by the search program 12 as various means (means relating to identifying various words, means relating to generating words, means relating to searching for data groups, etc.). The search program 12 will be explained in detail later, and first, the tables 13 to 16 and databases 17 and 18 included in the table DB 11 will be explained.

[0067] Figure 4 shows an example of part of the contents of the drug link table 13 stored in the table DB11 in the storage unit 10g. The drug link table 13 is a table that stores, for each of a plurality of drugs, side effect terms that indicate the side effects of the drug, in correspondence with the name of the drug (drug name), and the contents show the drug name corresponding to the side effect term. Specifically, the drug link table 13 is configured so that corresponding information such as the drug name, drug code, side effect severity, side effect occurrence frequency, side effect code, and side effect term is arranged in one row (horizontal row; the same applies below) for each drug.

[0068] As an example, for a drug named "TS-1 Combination Capsule T20," the following information is stored in a single line: drug code "4229101M1033" to identify the drug; severity of the drug's side effects "1"; frequency of mentions of the drug's side effects "H"; side effect code "000001" to identify the side effects; and side effect term "leukopenia" to indicate the drug's side effects.

[0069] The drug names stored in this drug link table 13 are generic names (commonly used names), and generic codes are also used for the drug codes. It is, of course, possible to use drug names and drug codes based on other standards; for example, it is conceivable to use drug names listed in the drug price standard for drug names, and information based on the drug price standard (which defines drugs that can be used in health insurance coverage) for drug codes. The side effect severity stored in the drug link table 13 is based on the description in the drug package insert and is indicated in two levels: "1" or "0" ("1" indicates a higher level of severity than "0"). Furthermore, the side effect occurrence frequency stored in the drug link table 13 is also based on the description in the drug package insert and is indicated in three levels: "H (high occurrence frequency)," "M (medium occurrence frequency)," and "L (low occurrence frequency)."

[0070] The side effect codes stored in the drug link table 13 are unique to this invention and are unrelated to drug price standards. Furthermore, the side effect terms stored in the drug link table 13 include terms indicating side effects used in MedDRA / J (ICH International Medical Dictionary, Japanese version), CTCAE (Common Terminology Criteria for Adverse Events), drug package inserts, interview forms, medical dictionaries, etc., as well as their synonyms.

[0071] Furthermore, such side effect terms include not only terms expressed as a single word such as "constipation" and "dizziness," but also terms formed by combining multiple words such as "leukopenia" and "low red blood cell count" ("leukopenia" is a side effect term formed by combining the two words "white blood cell" and "decrease," while "low red blood cell count" is a side effect term formed by combining the two words "red blood cell" and "low count"). The processing of the present invention targets side effect terms formed by combining multiple words. The contents of the drug link table 13 are updated each time a new drug is made available on the market.

[0072] In addition, since it is common for a single drug to cause multiple side effects, the drug link table 13 in Figure 4 also includes content that associates multiple different side effect terms with the same drug name.

[0073] FIG. 5 shows an example of part of the contents of the synonymous side effect table 14 stored in the table DB11 in the storage unit 10g. The synonymous side effect table 14 shows side effect terms that are synonymous with side effect terms, and these synonymous side effect terms include side effect terms formed by combining multiple words (referred to as word-combined side effect terms). The synonymous side effect table 14 of this embodiment shows side effect terms that are synonymous with side effect terms formed by combining multiple words among the side effect terms included in the drug link table 13 shown in FIG. 4, and stores side effect terms that are synonymous with each side effect term. Each of these side effect terms is assigned a side effect code that identifies it, and such side effect codes are assigned uniquely.

[0074] For example, for the side effect term "leukopenia (side effect code is 000001)," the synonymous side effect term "decreased WBC (side effect code is 000002)" is displayed, and for the side effect term "toxic epidermal necrolysis (side effect code is 000006)," three synonymous side effect terms are displayed: "toxic epidermal necrolysis (side effect code is 000007)," "Lyell's syndrome (side effect code is 000008)," and "TEN (side effect code is 000009)." Furthermore, the synonymous side effect table 14 in FIG. 5 displays the synonymous side effect terms "decreased red blood cells (side effect code is 000003)" and "decreased RBC (side effect code is 000103)" for the side effect term "low red blood cell count (side effect code is 000003)" (RBC stands for "red blood cell").

[0075] In this way, the synonymous side effect table 14 of this embodiment is configured so that side effect terms that are synonymous with a certain side effect term are shown in one line, and the synonymous side effect terms are not limited to just one word, but include side effect terms for which there are multiple (two or more) synonymous side effect terms, such as the example of the side effect term "toxic epidermal necrolysis" mentioned above.

[0076] Fig. 6 shows an example of part of the contents of the side effect word table 15 stored in the table DB 11 in the storage unit 10g. The side effect word table 15 shows the individual words used in each side effect term (word-combined side effect term) formed by combining multiple words among the side effect terms included in the synonymous side effect table 14 shown in Fig. 5, as well as the importance of each word in the side effect term and the order in which each word is combined as co-occurrence.

[0077] Specifically, the side effect word table 15 lists, for each side effect term, each word that constitutes the side effect term in the same row, indicating that each of these words is used in combination.

[0078] Furthermore, the side effect word table 15 assigns a level value to each word constituting a side effect term, indicating the degree of importance of that word in expressing the meaning of the side effect term, as the importance of each word in the side effect term. The most important word is assigned a value (level value) of "1" (level 1 indicates the most important word in expressing the meaning of the side effect term), the second most important word is assigned a value (level value) of "2," and so on for side effect terms consisting of three or more words. However, words with low semantic importance, such as particles (e.g., "na" and "no"), are assigned a value (level value) of "0" so that their values ​​are not counted. Furthermore, the side effect word table 15 indicates the order in which words are combined, i.e., the order in which words are arranged from left to right in a line, indicating the order in which they are used in combination within the side effect term.

[0079] For example, for the side effect term "leukopenia" with the side effect code "000001," side effect word table 15 indicates that the two words "leukocyte" and "decrease" are used in combination by placing the words "leukocyte" and "decrease" in the same row. Furthermore, in terms of semantic importance, side effect word table 15 indicates that the word "leukocyte" is at level 1 and the word "decrease" is at level 2. Furthermore, in terms of the combining order of each word, side effect word table 15 places the word "leukocyte" first from the left and the word "decrease" second among the words used in combination in the same row containing the side effect term "leukopenia," indicating that the combining order of "leukocyte" is first (first from the beginning of the word) and the combining order of "decrease" is second (second from the beginning of the word).

[0080] As another specific example, for the side effect term "severe bone marrow suppression" with side effect code "000005," side effect word table 15 places the words "severe," "of," "bone marrow," and "suppression" in the same row, thereby indicating that four words, "severe," "of," "bone marrow," and "suppression," are used in combination. Furthermore, in terms of semantic importance, side effect word table 15 indicates that the word "severe" is at level 3, "of" is at level 0, "bone marrow" is at level 1, and "suppression" is at level 2. Furthermore, in terms of the combining order of each word, side effect word table 15 places "severe" first from the left in the same row containing the side effect term "severe bone marrow suppression," followed by "of" second, "bone marrow" third, and "suppression" fourth, thereby indicating that the combining order (from the beginning of the word) of these words is "severe," "of" second, "bone marrow" third, and "suppression" fourth.

[0081] The side effect word table 15 of the above-mentioned arrangement configuration also shows the co-occurrence of each word used in combination. Co-occurrence means a collocation relationship in which two or more words appear in association with each other in the same sentence or word.

[0082] 8(a) shows an example of co-occurrence (e.g., the order of words) of words included in the side effect word table 15. Since "white blood cell" and "decrease" are words that make up the phrase "white blood cell decrease," they form a collocation that appears in relation to each other. By including the words "white blood cell" and "decrease" in the same row as the side effect term "white blood cell decrease," the side effect word table 15 shows that "decrease" co-occurs with "white blood cell." Similarly, since "WBC" and "decrease" are words that make up the phrase "WBC decrease," they form a collocation that appears in relation to each other. By including "WBC" and "decrease" in the same row as the side effect term "WBC decrease," the side effect word table 15 shows that "WBC" and "decrease" co-occur.

[0083] Fig. 7 shows an example of part of the contents of the synonymous word table 16 stored in the table DB11 in the storage unit 10g. The synonymous word table 16 indicates words that have a synonymous relationship with each of the words that make up each of the multiple side effect terms (word-combined side effect terms) included in the side effect word table 15 shown in Fig. 6. Specifically, the synonymous word table 16 indicates that each word included in the same line has a synonymous relationship, and each word is associated with a code (word code) that identifies the word and a level numerical value shown in the side effect word table 15 described above.

[0084] For example, synonym table 16 indicates that the word "white blood cell" is synonymous with the word "WBC" contained in the same row, and indicates that "white blood cell" has the word code "T000001" and is at level 1, and that "WBC" has the word code "T000002" and is at level 1. Synonym table 16 also indicates that the word "decrease" is synonymous with the word "decrease" contained in the same row, and indicates that "decrease" has the word code "T000010" and is at level 2, and that "decrease" has the word code "T000011" and is at level 2.

[0085] The user database 17 (see FIG. 3) stored in the storage unit 10g stores user information (such as name, telephone number, email address, date of birth, and password) related to registered users for each user ID, and the information stored in the user database 17 is confirmed during user authentication, etc. In this embodiment, an email address is used as the login ID for logging in to the information retrieval system 1.

[0086] The screen database 18 stored in the storage unit 10g stores various screen data corresponding to the screen contents (see FIGS. 10, 11, etc.) to be displayed on the terminal device 30 that has accessed the server device 10. The server device 10 reads out screen data corresponding to the process from the screen database 18, processes and generates the screen data to have predetermined contents, and transmits it to the terminal device 30.

[0087] Next, each process defined by the search program 12 (corresponding to a computer program according to the present invention) stored in the storage unit 10g will be described. The main processes defined by the search program 12 to be executed by the MPU 10a include user authentication, search for adverse drug reaction terms, identification of drug names based on search results, and information extraction from data sets based on search results. By executing each of these processes according to the definitions of the search program 12, the MPU 10a functions as various means, enabling ambiguous searches for the contents of the same adverse drug reaction using various notations.

[0088] First, as a user authentication process, when the server device 10 is accessed from a terminal device (for example, the terminal device 30), the MPU 10a performs a process of transmitting screen data corresponding to a login screen 41 as shown in Fig. 10(a) to the accessing terminal device 30. Then, in response to the transmission of this screen data, the server device 10 receives login information including a user ID and a password from the accessing terminal device 30, and the MPU 10a determines whether the received user ID and password are stored in the user database 17.

[0089] If it is found that the sent user ID and password are not stored in the user database 17, login is not permitted, and the MPU 10a transmits a login-prohibited notice to the accessing terminal device 30. On the other hand, if the user ID and password are stored, the MPU 10a stores in the RAM 10c the fact that the user of the accessed terminal device 30 is now logged in, and also reads out screen data corresponding to the search target confirmation screen 42 shown in Fig. 10(b) from the screen database 18 and transmits it to the accessing terminal device 30 together with a login completion notice.

[0090] The information retrieval system 1 of this embodiment is a system built specifically for each data group to be searched. Examples of data groups to be searched include big data 5, X Hospital's electronic medical record DB 6, and Y Pharmacy's electronic medical history DB 7, as shown in Figures 2(a)-(c). One information retrieval system 1 may be dedicated to searching big data 5, another information retrieval system 1 may be dedicated to searching X Hospital's electronic medical record DB 6, and yet another information retrieval system may be dedicated to searching Y Pharmacy's electronic medical history DB 7.

[0091] The search target confirmation screen 42 shown in FIG. 10(b) is displayed when big data 5 is set as a dedicated search target in each dedicated information retrieval system 1 as described above. The screen content presents the big data 5 to be searched and medical-related literature A, B, C, etc. contained in the big data 5, and also includes a selectable "Next" button 42a. The search target confirmation screen 42 is designed so that when a user's selection operation is accepted using the "Next" button 42a, a notification of the selection of the button 42a is sent to the server device 10. When the server device 10 receives a notification of the selection of the button 42 sent from the accessing terminal device 30, the server device 10 performs processing to display a search word setting screen on the terminal device 30 so that the user can set and input search words (side effect terms, etc.) on the accessing terminal device 30 (the server device 10 enters keyword search mode and proceeds with the processing).

[0092] 11 shows a search word setting screen 43 displayed on the terminal device 30. When the server device 10 receives a notification that the button 42 has been selected, the MPU 10a reads screen data corresponding to this screen from the screen database 18 and transmits it to the accessing terminal device 30, thereby displaying the search word setting screen 43 on the terminal device 30. After transmitting this screen data, the server device 10 waits for the search word (such as a side effect term for search) set and input on the search word setting screen 44 to be transmitted from the terminal device 30.

[0093] In this embodiment, a side effect term consisting of a combination of multiple words is used as a search word, but as will be described later, there is also a process for using a side effect term consisting of a single word as a search word. Furthermore, the search word setting screen 43 in Figure 11 is designed to allow setting and input of at least one side effect term, and also allows input of drug names as search words.

[0094] 11 is basically performed by the information retrieval system 1, which is specialized for searching data groups such as literature, such as big data 5, depending on the system specifications, but it may also be possible to allow the user to set whether they wish to use "keyword search" or "automatic search" even when searching literature or other types of information. In such a case, when the server device 10 receives a notification that the above-mentioned button 42 has been selected, the information retrieval system 1 performs processing to display a search type setting screen on the terminal device 30.

[0095] 12(a) shows a search type setting screen 44 displayed on the terminal device 30, which allows the user to set whether they wish to use "keyword search" or "automatic search." As described above, the server device 10 (MPU 10a), upon receiving the selection of the button 42, follows the rules of the search program 12 and reads out screen data corresponding to the search type setting screen 44 from the screen database 18 and transmits it to the accessing terminal device 30, whereby the search type setting screen 44 is displayed on the terminal device 30.

[0096] When the server device 10 receives a notification of "keyword search" from the accessing terminal device 30, the server device 10 enters keyword search mode, and when it receives a notification of "automatic search," the server device 10 enters automatic search mode. When the server device 10 enters keyword search mode, the MPU 10a performs a process of displaying the search word setting screen 43 shown in Fig. 11 on the accessing terminal device 30.

[0097] When the search word entered by the user on the search word setting screen 43 of Figure 11 is sent from the terminal device 30 to the server device 10, and the server device 10 (MPU 10a) receives the search word, it is temporarily stored in the RAM 10c, etc., and the MPU 10a performs a process as a side effect term identification means to identify word-combined side effect terms that become synonymous side effect terms for side effect terms (corresponding to side effect terms for search that are formed by combining multiple words) in the received search word based on the synonymous side effect table 14 shown in Figure 5 (or the synonymous side effect table 54 shown in Figure 20).

[0098] For example, if the received search word is a side effect term "leukopenia" (a search side effect term formed by combining multiple words), the MPU 10a identifies "WBC decrease" as a synonymous side effect term (word-combined side effect term) from the synonymous side effect table 14 in Figure 5. On the other hand, conversely to the above example, if the received search word is a side effect term "WBC decrease" (a search side effect term formed by combining multiple words), the MPU 10a identifies "leukopenia" as a synonymous side effect term (word-combined side effect term) from the synonymous side effect table 14. In essence, the MPU 10a identifies a row from the synonymous side effect table 14 in Figure 5 that includes the search side effect term of the received search word, and then performs a process of identifying side effect terms consisting of multiple words, other than the received search word, included in the identified row as synonymous word-combined side effect terms.

[0099] Next, the MPU 10a performs a process as a word identification means to identify the individual words used in the combination for each of the received side effect terms (side effect terms for search formed by combining multiple words) and synonymous side effect terms (synonymous word-combined side effect terms) identified from the synonymous side effect table 14, based on the side effect word table 15 shown in Figure 6.

[0100] For example, if the received side effect term (side effect term for search) is "leukopenia" and the side effect term (synonymous side effect term) identified from the synonymous side effect table 14 of Figure 5 is "decreased WBC", the MPU 10a performs processing to decompose "decreased white blood cells" into two words, "white blood cells" and "decreased", and to identify "decreased WBC" by decomposing it into two words, "WBC" and "decreased", based on the side effect word table 15 of Figure 6.

[0101] Then, the MPU 10a performs a process of identifying a synonymous word group for each of the decomposed (identified) individual words, based on the synonymous word table 16 shown in Fig. 7, as synonymous word group identification means. Continuing with the above example, for the two words "white blood cell" and "decrease" decomposed from "white blood cell decrease", and the two words "WBC" and "decrease" decomposed (identified) from "WBC decrease", since "white blood cell" and "WBC" are included in the same row in the synonymous word table 16, the MPU 10a identifies these words as a synonymous word group, and since "decrease" and "decrease" are included in a different row in the synonymous word table 16, the MPU 10a identifies these words as a synonymous word group (see also Fig. 8(b)).

[0102] Then, the MPU 10a performs a process as generation means to extract one word from each of the identified synonymous word groups and combine the extracted words in the combining order shown in the side effect word table 15 of Fig. 6 to generate side effect terms. At this time, the MPU 10a performs a process to combine the generated side effect terms so that they are written differently from the above-mentioned "side effect terms for search" and the above-mentioned "synonymous side effect terms" to generate "side effect terms with different spellings." To generate such "side effect terms with different spellings," the MPU 10a extracts words from each synonymous word group so that the combinations are different from the combinations of each word used to combine the above-mentioned "side effect terms for search" and the combinations of each word used to combine the above-mentioned "synonymous side effect terms."

[0103] Continuing with the example above, if the "side effect term for search" is "leukopenia" and the "synonymous side effect term (synonymous word combination side effect term)" is "decreased WBC," the words included in the first synonymous word group from the beginning are "white blood cell" and "WBC," and the words included in the second synonymous word group from the beginning are "decreased" and "decreased."

[0104] 8(c), for example, if "white blood cell" is extracted from the first synonymous word group, extracting "decreased" from the second synonymous word group will result in the same spelling as "white blood cell decrease" in the "side effect terms for search," so the MPU 10a extracts "decreased" from the second synonymous word group, thereby generating a side effect term "white blood cell decrease," which is a different spelling from "white blood cell decrease" and "WBC decrease." Furthermore, if the MPU 10a extracts "WBC" from the first synonymous word group, extracting "decreased" from the second synonymous word group will result in the same spelling as "WBC decrease" in the "synonymous side effect terms (synonymous word combination side effect terms)," so the MPU 10a extracts "decreased" from the second synonymous word group, thereby generating a side effect term "WBC decrease," which is a different spelling from "white blood cell decrease" and "WBC decrease."

[0105] In the above example, the first synonymous word group from the beginning includes the two words "white blood cell" and "WBC," and the second synonymous word group from the beginning includes the two words "decreased" and "decreased," so a total of four patterns of notation can be generated as side effect terms by combining these words. However, of these four patterns, "white blood cell decrease" and "decreased WBC" are already used as "side effect terms for search" and "synonymous side effect terms (synonymous word combination side effect terms)," so MPU10a generates the remaining "decreased white blood cell" and "decreased WBC" as side effect terms (side effect terms with different notation) that are notated differently from the "side effect terms for search" and "synonymous side effect terms (synonymous word combination side effect terms)."

[0106] Note that the generation of "variantly spelled side effect terms" is basically the same in other examples, but if there are multiple "synonymous side effect terms (synonymous word-combined side effect terms)" for the "search side effect term," the number of "variantly spelled side effect terms" generated may be greater than in the example described above. In the example described above, the "search side effect term" is one word (e.g., "leukopenia") and the "synonymous side effect term" is one word ("decreased WBC"), so two "variantly spelled side effect terms (decreased white blood cell count, decreased WBC)" are generated. However, if there are multiple "synonymous side effect terms (synonymous word-combined side effect terms)" (e.g., two "synonymous side effect terms (synonymous word-combined side effect terms)"), more than two "variantly spelled side effect terms" will be generated (e.g., if there are two "synonymous side effect terms (synonymous word-combined side effect terms)," six "variantly spelled side effect terms" may be generated).

[0107] Next, the MPU 10a performs a search process for a data group to be searched designated by the user (terminal device 30) using the above-mentioned "side effect terms for search," "synonymous side effect terms (side effect terms formed by combining synonymous words)," and the generated "side effect terms with different spellings" as search words. An example of a data group to be searched is big data 5, as described above. In this embodiment, the contents of such big data 5 are literature information (information accumulating the literature contents of medical-related literature A, B, C, etc.) as shown in Figure 2(a), and therefore include information on various drug names and side effect terms indicating the side effects of the drugs corresponding to those drug names.

[0108] The above search process determines whether Big Data 5 contains any of the search words, "search side effect terms," ​​"synonymous side effect terms (synonymous word combination side effect terms)," or "differently spelled side effect terms" (whether or not the search process matches).

[0109] If the search words for such a search process are the above-mentioned "leukopenia" as the "side effect term for search," "decreased WBC" as the "synonymous side effect term (synonymous word combination side effect term)," and "decreased leukopenia" and "decreased WBC" as "side effect terms with different spellings," when "decreased leukopenia" as the "side effect term for search" or "decreased WBC" as the "synonymous side effect term (synonymous word combination side effect term)" match (a hit), the search results will not be much different from those of the conventional technology. However, when the "side effect terms with different spellings" "decreased leukopenia" or "decreased WBC" match (a hit), these are not used as search words in the conventional technology. Therefore, the ability to perform search processing using such "side effect terms with different spellings" is an advantage of the present invention, and therefore, when "side effect terms with different spellings" are hit in the search process, it helps to prevent missed searches.

[0110] The MPU 10a then performs a process as drug name identification means to identify the drug name corresponding to the side effect term found in the search process described above, based on the drug link table 13 shown in Fig. 4. In the process of identifying the drug name from this drug link table 13, there are cases where the hit side effect term (the side effect term found in the search process) is included in the drug link table 13 and cases where it is not, and therefore the processing procedure of the MPU 10a for identifying the drug name differs in each case.

[0111] Continuing with the above example, the fact that different processing procedures occur in identifying such drug names will be explained. In this example, the drug link table 13 shown in Figure 4 includes "leukopenia" in the "side effect terms for search," but does not include "decreased WBC" in the "synonymous side effect terms (synonymous word combination side effect terms)," nor "decreased WBC" and "decreased WBC" in the "side effect terms with different spellings." If the side effect term hit in the search process for the data group is "decreased WBC" in the "side effect terms for search," "decreased WBC" is included in the drug link table 13, and therefore, as shown in Figure 4, the MPU 10a can identify the drug name "TS-1 Combination Capsule T20" that corresponds to "decreased WBC."

[0112] On the other hand, if the side effect term hit in the search process of the data group is "WBC decrease," which is a "synonymous side effect term (synonymous word combination side effect term)," the MPU 10a will first perform a process (detection process) to search for the hit "WBC decrease" in the drug link table 13 of Figure 4.However, in this example, since "WBC decrease" is not included in the drug link table 13, the result is that "WBC decrease" is not found (not detected).

[0113] If the result is that the search result is not found (not detected), the MPU 10a performs a process (detection process) of searching whether other search words used in the search process of the data group described above are included in the drug link table 13. In this example, in addition to "WBC decrease" among the "synonymous side effect terms (synonymous word combination side effect terms)," other search words were used, including "leukopenia" among the "side effect terms for search" and "leukopenia" and "WBC decrease" among the "side effect terms with different spellings," so the MPU 10a performs a process of detecting whether "leukopenia," "leukopenia," and "WBC decrease" are included in the drug link table 13. Then, in this example, because "leukopenia" is included in the drug link table 13, the MPU 10a identifies the drug name "TS-1 Combination Capsule T20" that corresponds to "leukopenia," and as a result, even if processing is performed using "WBC decrease," the drug name "TS-1 Combination Capsule T20" is ultimately identified.

[0114] Similarly, if the side effect term hit in the search process for the data group is "decreased white blood cell count" or "decreased WBC count" in the "side effect terms with different spellings," these side effect terms are not included in the drug link table 13 of Figure 4, so by performing the same process as in the case of "decreased WBC count" in the "synonymous side effect terms" described above, the MPU 10a will ultimately identify the drug name "TS-1 Combined Capsule T20" that corresponds to "decreased white blood cell count" based on the drug link table 13.

[0115] The method of identifying drug names, which includes these two different processing procedures, as a whole, is the process of identifying drug names corresponding to the side effect terms found in the data group search process from the drug link table 13. Note that in the drug link table 13 shown in Figure 4, only one drug name, "TS-1 Combination Capsule T20," corresponds to the side effect term "leukopenia," but an actual drug link table 13 will contain multiple (many) drug names corresponding to the side effect term "leukopenia."

[0116] Then, the MPU 10a performs an output process of the side effect terms found in the data group search process and the drug names identified in the drug link table 13. As a specific content of this output process, the MPU 10a generates screen data of a drug name list screen that shows the processing results as a list, and transmits the generated screen data to the terminal device 30 that is the access source.

[0117] At this time, the MPU 10a generates screen data configured to include a list that associates the side effect terms used to identify the drug name from the drug name link table 13 in Fig. 4 with the drug names identified from those side effect terms. When identifying a drug name from the drug name link table 13, if the side effect terms matched in the data group search process are not included in the drug link table 13 and the drug name is identified from the drug link table 13 using a side effect term written in a different way (a side effect term of another search word), the MPU 10a generates screen data including a list that associates the side effect terms used to identify the drug name in addition to the identified drug name and the side effect terms matched in the data group search process.

[0118] Furthermore, since it is generally expected that there will be multiple drug names identified from the above-mentioned drug name link table 13, when generating screen data for this drug list screen, multiple drug names are arranged in a list included in the screen data. At this time, the drug names are arranged in the list in an order such that drug names corresponding to adverse reaction terms with higher adverse reaction severity levels (drug names with "1" attached) shown in the drug link table 13 of Figure 4 are placed at the top (vertically higher in the list).

[0119] 13 shows an example of a drug name list screen 46 corresponding to the screen data generated by the MPU 10a. This drug name list screen 46 includes a correspondence table 20 that associates the side effect terms used in the identification process with the drug names identified by those side effect terms. The associated side effect terms and drug names are arranged in the same horizontal row, with the side effect terms (side effect terms shown in the side effect term column 20a) matched in the data group search process described above and the drug names corresponding to those side effect terms (drug names identified based on the drug link table 13). This allows the user to see at a glance what drugs are available for a particular side effect on the drug name list screen 46.

[0120] Furthermore, in this correspondence list 20, each drug name included in the drug name column 20b corresponding to the search target side effect term column 20a indicating the side effect term matched in the search process can be selected by a user operation on the terminal device 30 (multiple drug names can be selected), and the screen data of the drug name list screen 46 is designed so that when a drug name included in the drug name column 20b is selected and a data search button 46c included in the drug name list screen 46 is selected, a notification of the selected drug name is sent to the server device 10. Note that the drug name list screen 46 has a search setting side effect column 46d above the correspondence list 20, and the side effect term of the search word entered by the user on the search word setting screen 43 shown in Figure 11 is arranged in this search setting side effect column 46d, thereby allowing the user to confirm both the search word (side effect term) entered by the user and the content shown in the correspondence list 20 corresponding to the processing result.

[0121] When the server device 10 transmits the screen data of the drug name list screen 46, it receives notification of the drug name and the side effect term corresponding to that drug name from the accessing terminal device 30, thereby accepting information such as any drug name selected on the terminal device 30.

[0122] When the server device 10 (MPU 10a) receives and accepts a notification (a notification informing the user of the drug name, etc. selected by the user) sent from the terminal device 30 (corresponding to accepting the drug name as a drug name accepting means), the MPU 10a performs a process of searching and extracting information from the data group that lists both the side effect term that was found in the above-mentioned data group search process (the same term as the side effect term accepted from the terminal device 30) and the drug name accepted from the terminal device 30. Note that if multiple drug names are selected in the drug name column 20b on the drug name list screen 46, the above-mentioned extraction process is performed for each selected drug name (corresponding to the selected drug name) (information that lists both the selected drug name and the side effect term corresponding to the selected drug name is extracted for each selected drug name).

[0123] In the case of the information retrieval system 1 in which big data 5 shown in FIG. 2(a) is a data group to be searched, this big data 5 is a data group of literature information including medical-related literature A, B, C, etc. as shown in FIG. 2(a), so the MPU 10a searches for and extracts (detects) medical-related literature in which both the above-mentioned side effect terms and drug names are written from the big data 5. The MPU 10a then generates screen data for a search result screen that shows information about the extracted medical-related literature as a list, and performs a process of transmitting the generated screen data to the accessing terminal device 30. The generated screen data includes a list that lists information such as the literature titles, author names, publication dates, and side effect terms and drug names described in the extracted medical-related literature.

[0124] 14 shows an example of a literature search result screen 47 corresponding to screen data generated by the MPU 10a. This literature search result screen 47 includes a literature information list 21 that displays information indicating extracted medical-related literature (including medical-related papers), and the literature information list 21 includes a word column 21a that displays the words (side effect terms and drug names) used to extract the literature, a literature name column 21b that displays the extracted literature names (including the names of papers (paper names) in addition to the names of the literature), and a bibliographic information column 21c that displays bibliographic information about the extracted literature, and each literature name displayed in the literature name column 21b can be selected by operating the accessing terminal device 30.

[0125] When a document name in the document name column 21b is selected on the terminal device 30, the data corresponding to the selected document name is retrieved from the big data 5 and displayed on the terminal device 30. Each document name in the document name column 21b is embedded with link information (e.g., information specifying the access destination on the network, such as a URL) that provides an access destination for the document in question in the big data 5. Furthermore, the bibliographic information column 21c contains bibliographic information about the document, such as the publication date, author (writer), and publisher. The word column 21a contains adverse reaction terms and drug names related to the process leading up to the extraction of the document name in the document name column 21b on the same line. The document search result screen 47 includes a scroll bar 47a to the right of the document information list 21, as well as a back-to-top button 47b, a back-to-top button 47c, and a browse button 47d below the document information list 21. In addition, the literature search results screen 47 has a search setting side effect column 47e above the literature information list table 21, similar to the drug name list screen 46 in Figure 13, which displays the side effect terms of the search word entered by the user on the search word setting screen 43 in Figure 11.

[0126] The present invention is characterized by the fact that, after the above-mentioned search process for side effect terms, the literature search result screen 47 as shown in Figure 14 is finally displayed on the terminal device 30, thereby preventing missed searches and realizing efficient screening when searching for medicines corresponding to side effects and medical-related literature that lists them.

[0127] Note that the above description has been based on the case where a single search word (side effect term for search) is input by the user on the search word setting screen 43 in Figure 11 (for example, when a single side effect term, "leukopenia," is input), but it is also possible that a user may input multiple search words. When multiple search words are input in this way, the MPU 10a performs processing for each input search word, from the process of identifying "synonymous side effect terms" based on the synonymous side effect table 14 described above to the process of searching a data group using "side effect terms for search, synonyms, and different notations." In the subsequent search process, the process of identifying drug names corresponding to the matching side effect term from the drug link table 13 in Figure 4 is performed using an AND search-like process, thereby narrowing down the drug names based on the input of multiple search words.

[0128] For example, suppose two search words, a first search word (a side effect term for the first search) and a second search word (a side effect term for the second search), are entered on the search word setting screen 43 in Fig. 11. In this case, the MPU 10a first performs, for the first search word, a process of identifying "synonymous side effect terms" based on the above-mentioned synonymous side effect table 14, and a process of searching a data group using "side effect terms for search, synonymous, and different spellings," and also performs, for the second search word, a process of identifying "synonymous side effect terms" based on the above-mentioned synonymous side effect table 14, and a process of searching a data group using "side effect terms for search, synonymous, and different spellings."

[0129] Next, the MPU 10a performs a process of identifying drug names corresponding to the side effect terms found in this search process from the drug link table 13 shown in Fig. 4, where an AND search-like process is performed. Specifically, the MPU 10a identifies drug names corresponding to both the side effect terms found in the search process based on the first search word (if multiple side effect terms are found in the search process, any one of the matched side effect terms) and the side effect terms found in the search process based on the second search word (if multiple side effect terms are found in the search process, any one of the matched side effect terms) from the drug link table 13 (identifying drug names corresponding to the two side effect terms from the drug link table 13).

[0130] After identifying the drug names in this manner, as in the case described above, screen data for a drug name list screen (see Figure 13) is generated, which lists multiple side effect terms matched in the search process (side effect terms matched in the process for the first search word and side effect terms matched in the process for the second search word) and the drug names corresponding to those multiple side effect terms (drug names identified from the drug link table 13), and is sent to the terminal device 30 from which the access originated.

[0131] On the drug name list screen (see FIG. 13) displayed on the terminal device 30, the number of drug names included in the list is narrowed down by the AND search-like identification process described above, compared to when a drug name is identified by entering a single search word on the search word setting screen 43 of FIG. 11. This has the advantage of reducing the effort required for selecting a drug name (selecting a column containing the drug name) on the drug name list screen (see FIG. 13). When a drug name is selected on the terminal device 30 on this drug name list screen, the selection is transmitted to the server device 10, as in the above-described case, and the server device 10 performs an extraction process for information (such as literature information) from the data group, and finally, the search result screen (see FIG. 14) is displayed on the terminal device 30. Therefore, when multiple search words are entered on the search word setting screen 43 of FIG. 11, the number of drug names is narrowed down as described above, thereby improving the efficiency of screening work, etc.

[0132] In addition, in the search word setting screen 43 of Figure 11, in addition to the search word of the side effect term, the name of a drug can also be entered at will, and when a drug name is entered, the MPU 10a does not generate and output screen data corresponding to the drug name list screen 46 shown in Figure 13, but generates and outputs screen data corresponding to the literature search result screen 47 shown in Figure 14.

[0133] Specifically, when a search word for a side effect term and a drug name are entered on the search word setting screen 44 of Figure 12(a), the MPU 10a, as a search term receiving means, receives the entered search words (side effect terms for search and drug names for search) as in the above case, and performs processing from identifying "synonymous side effect terms" based on the above-mentioned synonymous side effect table 14 to searching data groups using "side effect terms for search, synonyms, and different notations."

[0134] In this case, if a side effect term is found by the search process of the data group, the MPU 10a then identifies the drug name corresponding to the side effect term found in the search process from the drug link table 13 of Fig. 4, but then does not generate or output screen data corresponding to the drug name list screen 46, but instead executes a process as drug name identification means to identify, among the identified drug names, the drug name that is the same as the drug name entered on the search word setting screen 44 (the accepted drug name for search).The MPU 10a then performs a process to detect and extract, from the data group, information (for example, pharmaceutical-related literature information) that lists both the side effect term found in the search process and the identified drug name that is the same as the drug name entered on the search word setting screen 44.

[0135] That is, in this case, since the search word setting screen 44 accepts input of the drug name desired by the user, there is no need to display the drug name list screen 46 as shown in Fig. 13 and select the drug name, and therefore, when the search process of the data group using "side effect terms for search, synonyms, and alternative notations" is completed, the drug names corresponding to the side effect terms found in the search process are identified from the drug link table 13 in Fig. 4, and the identified drug names are narrowed down to those that are the same as the drug name input on the search word setting screen 44. This narrows down the number of drug names used to extract information from the data group, making it possible to extract information from the data group efficiently.

[0136] If multiple drug names are entered on the search word setting screen 43 in Figure 11, the above-mentioned processes are performed for each entered drug name, and information containing both the side effect terms found in the search process and the identified same drug name is extracted from the data group.

[0137] For example, when two drug names, a first drug name and a second drug name, are entered, a process is performed to extract from the data set information about the first drug name that lists both the side effect term found in the search process and the same drug name as the identified first drug name, and a process is performed to extract from the data set information about the second drug name that lists both the side effect term found in the search process and the same drug name as the identified second drug name. Then, the MPU 10a generates screen data for a search result screen (see literature search result screen 47 in FIG. 14) that includes both pieces of information extracted by these two processes as a list, and transmits this data to the accessing terminal device 30. Through this process, the user can enter multiple desired drug names to obtain a final search result screen with more refined content.

[0138] The above describes the processing for keyword search. However, as mentioned above, when searching for data other than documents, the user can set whether they want a "keyword search" or an "automatic search" depending on the system specifications, etc. In this way, when the user can set the type of search, the search type setting screen 44 shown in FIG. 12(a) is displayed on the terminal device 30, and the user sets whether they want a "keyword search" or an "automatic search." Then, when the server device 10 (MPU 10a) receives a notification of "automatic search" after sending screen data corresponding to the search type setting screen 44 shown in FIG. 12(a) to the accessing terminal device 30, the processing of the MPU 10a enters automatic search mode.

[0139] When the automatic search mode is entered, the MPU 10a reads out screen data corresponding to the automatic search confirmation screen 45 as shown in Fig. 12(b) from the screen database 18 and transmits it to the accessing terminal device 30. If an "execute" notification is received along with the transmission of this screen data, the server device 10 (MPU 10a) proceeds with the automatic search process. In the automatic search process, the side effect terms included in the drug link table 13 in Fig. 4 are used as search words (side effect terms for search) to perform the process.

[0140] In the automatic search process of this embodiment, all side effect terms included in the drug link table 13 are used as search words for the search side effect terms. Therefore, if the drug link table 13 contains 1000 side effect terms, the MPU 10a performs, for each of those 1000 side effect terms, a process of identifying "synonymous side effect terms" based on the above-described synonymous side effect table 14, and a process of identifying from the drug link table 13 the drug names corresponding to the side effect terms found in the search process. Then, the MPU 10a generates image data of a drug name list screen including a list that associates the drug names identified in the above-described process of the 1000 side effect terms with the side effect terms found in the search process, and performs a process (output process) of transmitting the image data to the accessing terminal device 30.

[0141] The number of drug names included in the list on the drug list screen corresponding to the screen data sent in this manner will generally be much greater than the drug list screen (see Figure 13) when a single side effect term is entered on the search word setting screen 43 of Figure 11.However, even if a side effect term is not entered as a search word, the drug list screen can be obtained based on multiple side effect terms included in the drug link table 13, making the "automatic search" suitable for applications such as grasping the overall overview processing results.

[0142] In addition, after the output processing of the screen data of the drug name list screen is performed during the ``automatic search'' processing, it is the same as when a side effect term is entered in the keyword search mode described above (different from when a drug name is also entered in the keyword search mode), and when the server device (MPU10a) is notified of the selected drug name from the accessing terminal device 30, it detects (extracts) information (medical-related literature information, etc.) containing both the notified drug name and the side effect term found in the search processing from the data group, generates screen data corresponding to the search result screen, and performs processing (output processing) to send it to the accessing terminal device 30.

[0143] In addition, in the above, the data group to be searched was basically explained as big data 5 (corresponding to the data group including literature information such as medical-related literature A, B, and C shown in Figure 2(a)), but the processing content is basically the same even if a data group including other different information is set. However, the content of the search result screen that is finally output will depend on the type of data group that was set.

[0144] 15 shows an example of an electronic medical record search result screen 48 by the information retrieval system 1 when the search target is a data group including patient information (for example, X Hospital electronic medical record DB6 managed by the electronic medical record system of X Hospital). When the search target is a data group including information related to electronic medical records, there are no document names in the medical-related documents A, B, C, etc. included in the big data 5 described above, and instead, the patient name is always written in the patient information included in the data group.

[0145] 15 includes a patient information list 22, which includes a first column 22a (word column) showing words such as side effect terms and drug names used in the patient extraction process, a second column 22b (a column showing patient identification information, etc., or patient information column) showing information related to the patient's name, etc., and a third column 22c (medical information column) showing information related to the patient's medical care, etc. (information showing hospital visits, the name of the attending physician, etc.). Each piece of patient information included in the second column 22b showing the patient's name, etc., can be selected by operating the accessing terminal device 30, just like the document name column 21b showing the document name on the document search result screen 47 of FIG. 14. When selected, the terminal device 30 accesses the X Hospital electronic medical record DB6 and has embedded therein link information for the access destination so that the terminal device 30 can display the electronic medical record information (text information) of the selected patient. In addition, the various words such as side effect terms, drug names, and symptom terms shown in the first column 22a etc. are written in the patient's electronic medical record, and indicate the side effects and symptoms occurring in the patient, or the drugs prescribed to the patient.

[0146] As described above, in the case of an information retrieval system 1 in which the data group to be searched is an electronic medical record database, the MPU 10a generates screen data corresponding to the electronic medical record search result screen 48 shown in FIG. 15 and outputs the screen data to the accessing terminal device 30. The electronic medical record search result screen 48, output in this manner and displayed on the terminal device 30, also functions as a patient alert screen. Because the patient names included on the screen are related to the side effects corresponding to the side effect terms in the search word, this screen is ideal for hospital staff, etc., to search for patients associated with specific side effects. Other than the above-mentioned sections, the electronic medical record search result screen 48 has the same configuration as the literature search result screen 47 in FIG. 14. A scroll bar 48a is located to the right of the patient information list 22, and selectable back-one button 48b, back-to-top button 48c, and view button 48d are located below the patient information list 22. A search setting side effect field 48e is located above the patient information list 22, displaying the side effect terms of the search word entered by the user on the search word setting screen 43.

[0147] 16 shows an example of an electronic medical history search result screen 49 by the information retrieval system 1 when the search target is a data group including a different type of patient information (for example, the Y Pharmacy electronic medical history DB7 managed by the Y Pharmacy's electronic medical history system). When the search target is a data group including information related to electronic medical history, just as when the search target is a data group including information related to electronic medical records, the patient information included in the data group includes the patient's name. Therefore, the electronic medical history search result screen 49 has arranged thereon a patient information list 23 including a first column 23a (word column) showing words such as side effect terms and drug names used to extract the patient, a second column 23b (column showing patient identification information, patient information column) showing information related to the patient's name, etc., and a third column 23c (medical information) showing medical information related to the prescription of drugs, etc. (prescription information showing the name of the prescribing doctor, the name of the pharmacist, etc.). Therefore, when the electronic medical history DB is set as the search target, the MPU 10a generates screen data corresponding to the electronic medical history search result screen 49 of FIG. 16 and outputs it to the terminal device 30 that is the access source.

[0148] In this electronic medical history search result screen 49, each piece of patient information included in the second column 23b indicating the patient name, etc., can be selected by operating the accessing terminal device 30, and when selected, the terminal device 30 accesses the Y Pharmacy electronic medical history DB7 and embeds link information for the access destination so that the electronic medical history information (text information) of the selected patient can be displayed. When displayed on the terminal device 30, this electronic medical history search result screen 49 functions as an alert screen for patients, similar to the above-mentioned electronic medical record search result screen 48, and since the patient names included on the screen are related to side effects corresponding to the side effect terms in the search words, it is useful when pharmacy staff, etc., want to search for patients related to a specific side effect of a drug. In addition, the electronic medical history search results screen 49 has the same configuration as the literature search results screen 47 in Figure 14 or the electronic medical record search results screen 48 in Figure 15, except for the above-mentioned parts.A scroll bar 49a is located to the right of the patient information list 23, and a back one button 49b, a back to the beginning button 49c, and a view button 49d are selectably located below the patient information list 23.A search setting side effect field 49e is located above the patient information list 23, and shows the side effect terms of the search word entered by the user on the search word setting screen 43.

[0149] In this embodiment, the screen data of each of the above-mentioned screens is provided (output) to the terminal device 30 via a website. The server device 10 constructs a website for information search on the network, accepts access from the terminal device 30, etc., and provides the required screen data as appropriate according to the operation status of the accessing terminal device 30. Next, the accessing terminal device 30 will be described.

[0150] Fig. 9 is a block diagram showing the main internal configuration of the terminal device 30. A computer (such as a desktop or notebook personal computer) equipped with a communication function or the like can be applied as the terminal device 30 of this embodiment, and Fig. 9 shows the configuration when a computer is used. Note that, in addition to a computer, a tablet terminal, a smartphone, or the like can also be applied as the terminal device 30, and the main internal components related to the present invention when a tablet terminal, a smartphone, or the like is used are basically the same as those in the block diagram shown in Fig. 9, so Fig. 9 shows a block diagram common to devices that can be applied as the terminal device 30.

[0151] The terminal device 30 is configured by connecting various devices etc. to a CPU 30a that performs overall control and various processing via internal connection lines 30h. The various devices etc. include a communication unit 30b, a ROM 30c, a RAM 30d, an input interface 30e, a display output interface 30f, and a storage device 30g.

[0152] The communication unit 30b corresponds to a connection communication device (e.g., a LAN module) with the network NW, and enables communication with the server device 10, etc. by connecting to the network NW via a required communication device (not shown, such as a router), etc. The ROM 30c stores programs that define the basic processing contents of the CPU 30a, and the RAM 30d temporarily stores contents, files, etc. associated with the processing of the CPU 30a.

[0153] The input interface 30e is connected to a keyboard, mouse, etc. that accept operation instructions, etc. from the user of the terminal device 30, and performs processing to accept operation instructions, etc. from the user and transmit them to the CPU 30a. The display output interface 30f is connected to a display (display output device), and outputs content associated with processing by the CPU 30a to the display and presents it to the user (see each screen 40, etc. shown in FIG. 10, etc.).

[0154] The storage device 30g is a storage means configured by a hard disk drive (HDD) or a solid state drive (SSD), and stores a terminal OS 31, an information search program 32, etc. The terminal OS 31 corresponds to an operating system that defines the basic processing performed by the CPU 30a to make the terminal device 30 function as a computer.

[0155] The information search program 32 is a computer program with a browser function that enables browsing of websites, etc., and is an application program (app) that specifies the processing contents of the CPU 30a related to setting big data, DB, etc. as the data group to be searched based on user operation, and displaying and outputting a screen related to the information contained in the set big data, DB, etc. on the display.

[0156] When the information search program 32 of this embodiment is installed in the storage device 30g, a selectable icon corresponding to the information search program 32 is generated by the processing of the terminal OS 31 described above, and the icon is arranged on a home screen, a menu screen, etc. When the user performs an operation to select the icon, the information search program 32 is started, and the CPU 30a performs processing to access the server device 10 and display a login screen 41 shown in Fig. 10(a) on the display.

[0157] 10(a) has a user ID input field 41a, a password input field 41b, and a login button 41c arranged thereon, and when required information is input into the user ID input field 41a and the password input field 41b by a user operation, the login button 41c becomes active and can be selected. Then, when the login button 41 is selected with a user ID and a password input into the user ID input field 41a and the password input field 41b, the login operation is accepted, and the input information (login information indicating the user ID and password) in each input field 41a and 41b is sent to the server device 10. The login screen 41 is designed so that

[0158] If the terminal device 30 receives a login failure notification from the server device 10 in response to the transmission of the login information, the CPU 30a performs processing to display a login screen on the display indicating that login is not possible because the entered user ID or password is incorrect, so that the user can log in again. Furthermore, if the terminal device 30 receives a login completion notification and screen data from the server device 10 in response to the transmission of the login information, and the terminal device 30 is used in the information retrieval system 1 that searches big data 5 shown in FIG. 2(a) as a data group (when only keyword search is performed), the terminal device 30 (CPU 30a) generates a search target confirmation screen 42 shown in FIG. 10(b) and switches the display to the generated search target confirmation screen 42. When the user selects the “Next” button 42a on the search setting target confirmation screen 42, the CPU 30a transmits a notification indicating the selection of the button 42a to the server device 10.

[0159] When the terminal device 30 receives the next screen data in response to the notification that button 42a has been selected, the CPU 30a generates the search word setting screen 43 shown in Figure 11 based on the screen data and performs processing to switch the display from the search target confirmation screen 42 to the search word setting screen 43.

[0160] 11 includes two input fields, a side effect term input field 43a and a drug name input field 43b, as well as two buttons, a back button 43c and a search button 43d. The side effect term input field 43a is a field (required input field) for inputting at least one side effect term (in this embodiment, a search side effect term formed by combining multiple words) that will be used as a search word by operating the terminal device 30. The drug name input field 43b is an optional input field for inputting at least one drug name by operating the terminal device 30.

[0161] The back button 43c is a button that can be selected by operating the terminal device 30, and when the terminal device 30 accepts the selection operation of the back button 43c, it switches the display and performs a process of returning to the search target confirmation screen 42 of FIG. 10(b). The search button 43d is a button that can be selected when at least one side effect term is entered in the side effect term input field 43a, and when the terminal device 30 accepts the input of a side effect term and accepts the selection operation of the search button 43d, it performs a process of transmitting all of the side effect terms accepted in the side effect term input field 43a to the server device 10. At this time, if the terminal device 30 accepts the input of a drug name in the drug name input field 43b, it also performs a process of transmitting all of the entered drug names to the server device 10.

[0162] Furthermore, when the terminal device 30 is used in an information retrieval system 1 in which the data group to be searched is capable of both keyword search and automatic search, when the "Next" button 42a on the search target confirmation screen 42 shown in Figure 10(b) is selected and a notification of the selection of button 42a is sent to the server device 10, the terminal device 30 receives screen data corresponding to the search type setting screen 44 of Figure 12(a), and based on that screen data, the CPU 30a generates the search type setting screen 44 and performs processing to switch the display from the search target confirmation screen 42 to the search type setting screen 44.

[0163] 12(a) is a screen for setting whether to perform a search using "keyword search" or "automatic search" as the type of search to be performed in the information search service according to the present invention, and includes a keyword search button 44a and an automatic search button 44b that can be selected. When the keyword search button 44a is selected, the terminal device 30 (CPU 30a) performs processing to transmit a notification of "keyword search" to the server device 10, and when the automatic search button 44b is selected, the terminal device 30 (CPU 30a) performs processing to transmit a notification of "automatic search" to the server device 10.

[0164] When a "keyword search" notification is sent, the process is the same as when displaying the search word setting screen 43 in Figure 11 described above, and screen data corresponding to the search word setting screen 43 is sent from the server device 10. When the terminal device 30 receives this screen data, the terminal device 30 (CPU 30a) generates the search word setting screen 43 based on the received screen data and performs a process to switch the display from the search type setting screen 44 to the search word setting screen 43.

[0165] Furthermore, when an "automatic search" notification is sent, screen data corresponding to the automatic search confirmation screen 45 is sent from the server device 10. When the terminal device 30 receives the screen data, the terminal device 30 (CPU 30a) generates the automatic search confirmation screen 45 based on the received screen data and performs processing to switch the display from the search type setting screen 44 to the automatic search confirmation screen 45 (see Figure 12(b)).

[0166] 12(b) is a screen for confirming with the user whether to execute an automatic search, and includes a text message asking the user to confirm whether to execute an automatic search, such as "Are you sure you want to execute an automatic search?", as well as two buttons, a back button 45a and an execute button 45b. Because automatic searches impose a large processing load on the server device 10, this automatic search confirmation screen 45 is presented to the user to request the user's confirmation of whether to execute an automatic search.

[0167] The back button 45a is a button that can be selected by operating the terminal device 30, similar to the back button 44c on the above-described search word setting screen 44, and when the terminal device 30 accepts the selection operation, it switches the display and performs processing to return to the search target confirmation screen 42 of Fig. 10(b). Furthermore, the execute button 45b is a selectable button for issuing an instruction to the server device 10 to execute an automatic search, and when the terminal device 30 accepts the selection operation of the execute button 44, it performs processing to send a notification of the instruction to execute an automatic search to the server device 10.

[0168] 13 shows an example of a screen content displayed on the terminal device 30 when a side effect term (e.g., leukopenia) is entered in the side effect term input field 43a of the search word setting screen 43 (see FIG. 11) and the search button 43d is selected. Screen data corresponding to such a drug name list screen 46 is sent from the server device 10 to the terminal device 30 as a processing result in the server device 10 when the terminal device 30 transmits the side effect term for search entered in the side effect term input field 43a to the server device 10. The terminal device 30 generates the drug name list screen 46 based on the screen data and performs processing to display it on the display (the terminal device 30 switches the display from the search word setting screen 43 to the drug name list screen 46).

[0169] 11, when a drug name is entered in the drug name input field 43b in addition to entering a side effect term and transmitted to the server device 10, the drug names are narrowed down by the drug name input, and therefore the display of the drug name list screen 46 in Fig. 13 is omitted. On the other hand, when the above-mentioned automatic search mode is entered, the drug name list screen 46 in Fig. 13 is displayed, but in this case, it is displayed after switching from the display of the automatic search confirmation screen 45 in Fig. 12(b).

[0170] As described above, this drug name list screen 46 includes a correspondence list 20 in which side effect terms and drug names are associated with each other. If the number of side effect terms and drug names in the correspondence list 20 is too large for the size of the display, they cannot be displayed all at once. Therefore, a scroll bar 46a is provided at the right end of the correspondence list 20, and by operating this scroll bar 46a up and down, the entire contents of the correspondence list 20 can be viewed.

[0171] The correspondence list 20 is composed of multiple rows, each row including a side effect term column 20a and a drug name column 20b, and the side effect terms placed in the side effect term column 20a of each row correspond to the drug names placed in the drug name column 20b.

[0172] As described above, since one side effect term may correspond to multiple drugs, the screen data received from the server device 10 may include a configuration in which multiple drug names are associated with one side effect term, and may also include a configuration in which the drug name identified from the drug link table 13 is associated with the side effect term found in the data group search process, as well as the side effect term used to identify the drug name.Therefore, the correspondence list 20 on the drug name list screen 46 may include a single drug name associated with multiple side effect terms indicating the same side effect. Furthermore, since some drugs cause multiple different side effects, in the case of such drug names, the correspondence list 20 may have a list of side effect terms with different contents associated with them in different rows.

[0173] By displaying the drug name list screen 46 configured as described above on the terminal device 30 and presenting it to the user, the user can check at a glance which drugs correspond to each side effect by looking at the correspondence list 20 in the drug name list screen 46. Furthermore, there is an advantage in that the user can easily check what drugs correspond to the side effect term entered on the search word setting screen 44 shown in Figure 12(a).

[0174] Furthermore, the drug name list screen 46 includes a text notation above the correspondence list 20 that reads "Please select the drug name to narrow down," and each of the individual drug names shown in the drug name column 20b in the correspondence list 20 can be selected by operating the terminal device 30. Note that when a selection operation is performed on a selected drug name, the selected state is canceled.

[0175] Furthermore, the drug name list screen 46 has two buttons, a back button 46b and a data search button 46c, arranged below the correspondence table 20. The back button 46b is a button that can be selected by operating the terminal device 30, and when the selection operation is accepted, the terminal device 30 switches the display and performs processing to return to the search word setting screen 43 of Fig. 11. The data search button 46c is a button that becomes selectable when at least one drug name is selected in the drug name column 20b in the correspondence table 20, and when the terminal device 30 accepts the selection operation of the data search button 46c, the MPU 10a performs processing to send a notification of the drug name (and the side effect term associated with that drug name) that is selected in the correspondence table 20 to the server device 10.

[0176] The literature search result screen 47 shown in Fig. 14 is an example of a screen showing the final search results (an example in which the data group to be searched is medical-related literature such as big data 5), ​​and there are two patterns for reaching the display of this screen. The first pattern is when the data search button 46c is selected on the drug name list screen 46 in Fig. 13 described above, and a notification of the drug name that has been selected in the correspondence list 20 is sent to the server device 10, and the terminal device 30 generates the literature search result screen 47 based on the screen data sent from the server device 10 and displays it on the display.

[0177] As described above, the literature search result screen 47 includes a literature information list 21 that displays information indicating medical-related literature extracted by processing in the server device 10, and the literature information list 21 has a word column 21a that displays drug names and side effect terms related to the literature name extraction processing, a literature name column 21b that displays literature names, a bibliographic information column 21c that displays bibliographic information about the literature, etc. Note that if the number of literature names and literature information displayed in the literature information list 21 is too large for the size of the display, they cannot be displayed at once. Therefore, a scroll bar 47a is provided on the right side of the literature information list 21, and the contents of the literature information list 21 can be checked in their entirety by operating this scroll bar 47a up and down.

[0178] The document information list 21 is composed of multiple lines, each containing various information shown in a word column 21a, a document name column 21b, and a bibliographic information column 21c, with each line containing corresponding information. Multiple variations of side effect terms may be used to identify a single drug name. In such cases, multiple variations of side effect terms are shown in the word column 21a. By including a document information list 21 with this content on the document search result screen 47, the user can see at a glance the drug names and side effects of those drug names listed in each document shown in the document information list 21.

[0179] Furthermore, the document search result screen 47 includes a text notation above the document information list 21 that reads, "Please select the document name whose contents you wish to check," and any one of the individual document names in the document name column 21b of the document information list 21 can be selected by operating the terminal device 30 (only one document name can be selected at a time). Note that if a selection operation is performed on a document name that is in a selected state, the selected state is cancelled.

[0180] The literature search result screen 47 has three selectable buttons arranged below the literature information list 21: a back one button 47b, a back to beginning button 47c, and a browse button 47d. When a selection operation of the back one button 47b is accepted, the terminal device 30 performs a process to return the display to the previous screen (in the case of the first display pattern, the display is switched back to the drug name list screen 46 of FIG. 13). When a selection operation of the back to beginning button 47c is accepted, the terminal device 30 performs a process to return the display to the search target confirmation screen 42 of FIG. 10(b).

[0181] The view button 47d is a button that becomes selectable when one of the document names shown in the document name column 21a in the document information list 21 is selected, and is used to view the contents of the medical-related document of the document name shown in the selected document name column 21a. Specifically, when the selection operation of the view button 47d is accepted, the terminal device 30 performs access processing based on the link information embedded in the document name that is selected in the document information list 21, acquires information on the target document (information corresponding to the contents of the medical-related document), and displays it on the display.

[0182] The second pattern that leads to the display of the literature search result screen 47 in Figure 14 is when, on the terminal device 30, a drug name is entered in addition to a side effect term on the search word setting screen 43 in Figure 11, and the search button 43d is selected, and then, in response to sending a notification of the entered side effect term and drug name to the server device 10, the terminal device 30 generates the literature search result screen 47 based on the screen data sent from the server device 10 and displays it on the display.

[0183] In this case, the server device 10 performs a literature extraction process from the data group based on the drug names, etc. included in the notification sent from the terminal device 30, generates screen data for a literature search result screen 47 including a list of literature information (literature information list 21) according to the results of the extraction process, and transmits it to the terminal device 30.The terminal device 30 receives this screen data, generates the literature search result screen 47, and switches the display from the search word setting screen 43 of Figure 11 to the literature search result screen 47 of Figure 14.

[0184] The literature search result screen 47 displayed in this second pattern has the same specifications as the literature search result screen 47 displayed in the first pattern described above, but the content and number of information about the literature (medical-related literature) included in the literature information list 21 will vary depending on the number of side effect terms and drug names entered in the search word setting screen 44. Below, we will explain, based on flowcharts 1-13 shown in Figures 17-19, how the individual explanations about the server device 10 and terminal device 30 described above are executed in cooperation within a series of processing contents related to the information search service processing.

[0185] The flowchart 1-3 shown in Figures 17-19 illustrates the main processing steps of the side effect information search method according to the first embodiment of the present invention. The server device 10 and terminal device 30 perform various processes, including a search process for a data group, to identify a drug name and ultimately display and output a search result screen (see Figures 14-16, etc.). The processing flow of the side effect information search method will be explained below in accordance with the flowchart 1-3. The flowchart 1 in Figure 17 illustrates the processing steps for a system specification in which the user can select either "keyword search" or "automatic search." The flowchart begins when the terminal device 30 launches the information search program 32 and displays the login screen 41 shown in Figure 10(a), completing login. The flowchart 1-3 omits the process of the server device 10 transmitting screen data for each screen displayed on the terminal device 30.

[0186] First, in the first flowchart, upon completion of login, the terminal device 30 (equivalent to the CPU 30a; the same applies below) performs a process of displaying a search target confirmation screen 42 shown in Fig. 10(b) (S1). On this search target confirmation screen 42, the terminal device 30 determines whether or not the "Next" button 42a has been selected and confirmation has been completed (S2). If the button 42a has not been selected (S2: NO), the terminal device 30 enters a state of waiting for a user selection operation. If the user has selected the button 42a (S2: YES), the terminal device 30 switches the display content to the search type setting screen 44 shown in Fig. 12(a) (S3).

[0187] The terminal device 30 determines whether a selection operation of the "Keyword Search" or "Automatic Search" button 44a, 44b has been accepted on the search type setting screen 44 (S4). If a selection of "Keyword Search" has been accepted (S4: Keyword Search), the terminal device 30 transmits a notification of "Keyword Search" to the server device 10, and accordingly switches the display content to the search word setting screen 43 shown in Fig. 11 (S7).

[0188] On this search word setting screen 43, the terminal device 30 determines whether or not a search word (a side effect term for searching made up of a combination of multiple words must be entered, and the name of the drug for searching is optional) has been entered and a selection operation of the search button 43d has been accepted (S8), and if a selection operation of the search button 44d has not been accepted (S8: NO), it determines whether or not a selection operation of the back button 43c has been accepted (S9). If a selection operation of the back button 43c has not been accepted (S9: NO), the process returns to the step S8, but if a selection operation of the back button 43c has been accepted (S9: YES), the process returns to the step S3, and the display is switched to the search type setting screen 44, allowing the user to select the search type (keyword search or automatic search) again.

[0189] At step S8, if a search word is input and the search button 43d is selected (S8: YES), the terminal device 30 performs processing to send a notification informing the server device 10 of the input search word (S10). After this, the processing of the terminal device 30 proceeds to step S21 of the second flowchart in Fig. 18 if the input search word is only a side effect term, and proceeds to step S41 of the third flowchart in Fig. 19 if the input search word includes a drug name in addition to a side effect term.

[0190] 12(a) displayed on the display (S4: automatic search), the display is switched to the automatic search confirmation screen 45 shown in FIG. 12(b) (S5), and the terminal device 30 determines whether the selection operation of the back button 45a or the execute button 45b has been accepted (S6). If the selection operation of the back button 45a has been accepted (S6: back), the process returns to step S3, the display is switched to the search type setting screen 44, and the terminal device 30 is again in a state where the search type (keyword search or automatic search) can be selected.

[0191] If the selection operation of the execute button 45b is accepted at the S6 stage (S6: Execute), the terminal device 30 performs a process of sending a notification of "automatic search" to the server device 10, and then the process proceeds to the S21 stage of the second flowchart in Figure 18.

[0192] Meanwhile, in the first flowchart, the server device 10 first determines whether or not it has received a notification of the search type selected by the user (S12). If the notification of the search type is not received at step S12 (S12: NO), the server device 10 waits for notification of the search type. If the notification of the search target is received (S12: YES), the server device 10 determines whether or not the received search type is a "keyword search" (S13). If the received search type is not a "keyword search" (S13: NO), the server device 10 sets the processing to an automatic search mode (S14). Subsequent processing in the server device 10 proceeds in the automatic search mode, and the processing of the server device 10 thereafter proceeds to step S31 of the second flowchart shown in FIG. 18.

[0193] On the other hand, if the received search type is a "keyword search" notification (S13: YES), the processing of the server device 10 is set to keyword search mode (S16), and subsequent processing in the server device 10 proceeds in keyword search mode.

[0194] Furthermore, when the keyword search mode is entered at step S16, the server device 10 determines whether or not it has received a search word sent from the terminal device 30 (S17). If it has not received a search word (S17: NO), it enters a state of waiting for reception of a search word. If it has received a search word (S17: YES), it temporarily stores the received search word in RAM 10c or the like, and determines whether or not the received search word includes a drug name (S18).

[0195] If the drug name is included (S18: YES), the MPU 10a has accepted the side effect term and the drug name for the search as a search term accepting means, and the processing of the server device 10 then proceeds to step S51 of the third flowchart shown in Fig. 19. If the drug name is not included (S18: NO), the processing of the server device 10 then proceeds to step S31 of the second flowchart shown in Fig. 18.

[0196] The second flowchart in Figure 18 shows the processing contents for two cases: when the search word does not include a drug name (when the search word is only a side effect term) in keyword search mode, and when it is in automatic search mode. First, the former case will be explained. In this second flowchart, the server device 10 (similar to the MPU 10a, and the same applies below) identifies "synonymous side effect terms (synonymous word-combined side effect terms)" for the side effect terms (search side effect terms formed by combining multiple words) of the search word accepted by notification (reception) based on the synonymous side effect table 14 in Figure 5 (S31).

[0197] Then, the server device 10 identifies each of the "notified side effect term (side effect term for search formed by combining multiple words)" and the "synonymous side effect term (side effect term formed by combining synonymous words)" identified in step S31 by breaking them down into individual words used for combining, based on the side effect word table 15 of Fig. 6 (S32). Furthermore, for each of the identified individual words, the server device 10 identifies a synonymous word group showing a synonymous relationship, based on the synonym word table 16 of Fig. 7 (S33), extracts one word for each identified synonymous word group, and combines the extracted words in the combining order shown in the side effect word table 15 of Fig. 6, combining them so that they are written differently from the "notified side effect term (side effect term for search)" and the "synonymous side effect term" identified in step S31, to generate a "side effect term with a different spelling" (S34).

[0198] Then, the server device 10 performs a search process on the data group to be searched (e.g., big data 5) using the "notified side effect term (side effect term for search)," the "synonymous side effect term (synonymous word combination side effect term)" identified in step S31, and the "differently spelled side effect term" identified in step S34 as search words (S35). Then, when the server device 10 finds a side effect term in the search process in step S35, it temporarily stores the hit side effect term in RAM 10c and identifies the drug name corresponding to the hit side effect term from the drug link table 13 in Fig. 4 (S36).

[0199] In addition, when multiple side effect terms are input as search words and these multiple side effect terms are notified to the server device 10, the server device will perform processing from steps S31 to S36 in the second flowchart of Figure 18 described above for each of the multiple notified side effect terms.

[0200] Then, the server device 10 performs an output process to generate screen data for a drug name list screen 46 (see Figure 13) including a list 20 that associates the side effect terms found in the search process of S35 with the drug names identified in the S36 stage, and transmits the screen data to the terminal device 30 (S37).

[0201] In addition, if multiple side effect terms are entered as search words, and if a side effect term is hit in each process associated with the multiple side effect terms entered in the search process of S35, then multiple side effect terms will be hit.In this case, when identifying the drug name from the drug link table 13 at step S36, the drug name will be identified by an AND search-like process of the multiple side effect terms that were hit (the AND search-like process identifies the drug name that corresponds to all of the multiple side effect terms that were hit (matched) in the search process of S35).This makes it possible to narrow down the drug names by entering multiple side effect terms.

[0202] Furthermore, if multiple drug names are identified in step S36, the order in which the drug names are placed in the correspondence list 20 included in the screen data of the drug name list screen 46 generated in step S37 is such that the drug names associated with side effect terms with higher side effect severity levels (those marked with "1") shown in the drug link table 13 of Figure 4 are placed at the top (vertically higher in the list 20) (the same applies to the automatic search mode described below).

[0203] If the terminal device 30 receives input of only side effect terms as search words at step S8 of the first flowchart in Fig. 17, it proceeds to step S21 of the second flowchart in Fig. 18 via step S10, where it determines whether or not it has received screen data corresponding to the drug name list screen 46 sent from the server device 10. If it has not received the screen data (S21: NO), it enters a state of waiting for reception. On the other hand, if it has received the screen data (S21: YES), the terminal device 30 uses the received screen data to generate the drug name list screen 46 of Fig. 13 based on the screen data and displays it on the display (S22).

[0204] Next, the terminal device 30 determines whether the selection of the back button 46b has been accepted on the displayed drug name list screen 46, or whether the selection of the data search button 46c has been accepted in the correspondence list 20 on the drug name list screen 46 while a drug name included in the drug name column 20b is selected (S23). If the selection operation of the back button 46b has been accepted (S23: back button), the process returns to step S7 of the first flowchart in FIG. 17, and the screen display is switched back to the search word setting screen 43 in FIG. 11. This allows the user to check the drug name list screen 46 displayed in step S22, and if the screen content does not include the expected drug name, for example, the user is provided with a situation in which they can re-enter the search word, allowing the user to smoothly perform a re-search process using the re-entered search word.

[0205] Also, at the step S23, if the selection of the data search button 46c is received while any drug name is selected (S23: Data search), the terminal device 30 sends a notification to the server device 10 to inform the server device 10 of the drug name that has been selected in the correspondence list 20 on the drug name list screen 46 (any drug name selected from the drug names included in the correspondence list 20) and the side effect terms, etc. associated with that drug name (S24).

[0206] After transmitting the screen data in step S37, the server device 10 determines whether or not it has received a notification of the drug name, etc. sent from the terminal device 30 (S38), and if it has not received it (S38: NO), it enters a state of waiting for reception. Furthermore, if it has received a notification of the drug name, etc. (S38: YES), it is considered to have accepted the drug name, etc. included in the received notification, and searches the data group (e.g., big data 5) to be searched for in step S12 using the accepted drug name and side effect term (the side effect term matched in the search process in step S35), and extracts (detects) information in which both the drug name and side effect term are written (e.g., medical literature information if the search target is big data 5) (S39).

[0207] Then, the server device 10 generates screen data (screen data of search results) corresponding to a search result screen (for example, literature search result screen 47 shown in FIG. 14) that displays the extracted information (for example, medical literature information) as a list, and transmits it to the terminal device 30 (S40).

[0208] Meanwhile, after sending notification of the drug name, etc. at step S24, the terminal device 30 determines whether or not it has received the screen data of the search results sent from the server device 10 (S25), and if it has not received it (S25: NO), it enters a state of waiting for reception. Also, if it has received the screen data of the search results (S25: YES), the terminal device 30 generates a search result screen (for example, the literature search result screen 47 in FIG. 14) based on the received screen data, and switches the display to show the generated search result screen (S26).

[0209] For example, in the case of the literature search result screen 47 in Figure 14, the search result screen displayed in this manner allows the user to check information (information extracted at step S39) on literature (medical-related literature) that contains both the specified drug name and side effect term using the literature information list 21 included in the literature search result screen 47, and by selecting the desired literature name from the literature information list 21, the contents of the literature can also be checked, as described above, which can be useful for performing screening work regarding side effects, etc.

[0210] If the data group to be searched is a hospital medical record DB (for example, the X Hospital electronic medical record DB6 shown in FIG. 2(b)), the server device 10 performs a search process on the X Hospital electronic medical record DB6 at step S35 of the second flowchart in FIG. 18, and also searches for and extracts (detects) patient information from the X Hospital electronic medical record DB6 at step S39. Then, at step S40, the screen data generated by the server device 10 also becomes screen data corresponding to the electronic medical record search result screen 48 in FIG.

[0211] As a result, at step S26, the screen displayed on the display of the terminal device 30 also becomes the electronic medical record search result screen 48 shown in Fig. 15, which is useful when a user (e.g., a medical professional such as a doctor or nurse) wants to look up patients associated with a certain drug or its side effects. In particular, when looking up patients associated with side effects, the electronic medical record search result screen 48 in Fig. 15 shows the patient name, and therefore also functions as an alert screen for patients who may experience side effects, making it easier for medical professionals to respond to patients quickly.

[0212] Furthermore, if the data group to be searched is a pharmacy's electronic medical history DB (for example, the Y Pharmacy electronic medical history DB7 shown in Figure 2(c)), the server device 10 will perform a search process on the Y Pharmacy electronic medical history DB7 at step S35 of the second flowchart in Figure 18, and will also search for and extract (detect) patient information from the Y Pharmacy electronic medical history DB7 at step S39, and the screen data generated at step S40 will also be screen data corresponding to the electronic medical history search result screen 49 in Figure 16.

[0213] As a result, at step S26, the screen displayed on the display of the terminal device 30 also becomes the electronic medication history search result screen 49 shown in Fig. 16, which is also useful when a user (e.g., a pharmacist) wants to search for patients related to a certain drug or its side effects. Therefore, in particular when searching for patients related to side effects, like the electronic medical record search result screen 48 in Fig. 16 described above, this electronic medication history search result screen 49 also functions as an alert screen that shows patients who are the target of an alert regarding the side effect term in the search word.

[0214] Depending on the side effect term entered by the user, there may be cases where the search process at step S35 of the second flowchart in Figure 18 does not result in a hit (no match in the search process). Although the process for such a case where there is no hit is omitted in the second flowchart, at this stage the server device 10 transmits screen data indicating that there are no hits to the terminal device 30. Upon receiving this screen data, the terminal device 30 switches the display to a screen indicating that there are no hits (a screen including a selectable confirmation button). Then, upon receiving a selection operation of the confirmation button on this screen, the process returns to step S7, and the search word setting screen 43 of Figure 11 is displayed again, providing the user with an opportunity to enter another search word.

[0215] Furthermore, depending on the drug name selected by the user at step S23 in the second flowchart of Figure 18, there may be cases where the information extraction (detection) process from the data group at step S39 does not find any relevant information. Although the process for such a case is also omitted from the second flowchart, at this stage, the server device 10 transmits screen data indicating that there were no hits to the terminal device 30. Upon receiving this screen data, the terminal device 30 switches the display to a screen indicating that there were no hits (a screen including a selectable confirmation button). In this case, if the terminal device 30 receives a selection operation on the confirmation button on the screen indicating that there were no hits, the process returns to step S22, and the drug name list screen 46 of Figure 13 is redisplayed, providing the user with an opportunity to select another drug name.

[0216] The third flowchart in Figure 19 shows the processing when the search word entered by the user includes a drug name, i.e., when a drug name is set as a search word in addition to a side effect term. In this third flowchart, the server device 10 identifies synonymous side effect terms (side effect terms for search) included in the received search word based on the synonymous side effect table 14 in Figure 5 (S51), similar to step S31 in the second flowchart in Figure 18 described above.

[0217] The processing of the server device 10 from steps S52 to S55 in the third flowchart of Figure 19 is similar to the processing from steps S32 to S35 in the second flowchart of Figure 18 described above, and at step S55, the server device 10 performs a search process on the data group to be set (e.g., big data 5) using side effect terms that will be search words ("side effect terms for search," "synonymous side effect terms," ​​and "side effect terms with different spellings").

[0218] Furthermore, the server device 10 identifies the drug names corresponding to the side effect terms hit (matched) in the search process at step S55 from the drug link table 13 of Fig. 4 (S56). Then, the server device 10 identifies the drug names that are the same (same spelling) as the received and accepted drug name (the drug name for search entered by the user) from among the drug names identified at step S56 (S57), and searches for and extracts (detects) information (e.g., information on medical-related literature) that describes both the side effect terms hit in the search process at step S55 and the drug name identified at step S57 from the data group to be searched (e.g., big data 5) (S58).

[0219] If the data group to be searched is big data 5 storing medical-related literature information, literature information (medical-related literature) that lists both the above side effect terms and drug names will be searched for and extracted at this step S58. Also, if the data group to be searched is a hospital's electronic medical record DB or a pharmacy's electronic medication history DB that stores patient information (patient information that lists patient names, etc.), patient information that lists both the above side effect terms and drug names will be searched for and extracted at this step S58.

[0220] Finally, similar to step S40 of the second flowchart in Figure 18 described above, the server device 10 generates screen data (screen data of search results) corresponding to a search result screen (e.g., literature search result screen 47 shown in Figure 14) that displays the extracted information (e.g., information related to medical-related literature) as a list, and transmits it to the terminal device 30 (S59).

[0221] On the other hand, when the terminal device 30 receives a side effect term and a drug name as search words input by the user, the process proceeds to step S41 of the third flowchart in Fig. 19 after processing step S10 of the first flowchart in Fig. 17, where it determines whether or not it has received screen data of the search results sent from the server device 10. If it has not received the screen data (S41: NO), it enters a state of waiting for reception, and if it has received the screen data of the search results (S41: YES), the terminal device 30 generates a search result screen (for example, the literature search result screen 47 of Fig. 14) based on the received screen data, similar to step S26 of the second flowchart in Fig. 18 described above, and switches the display to display the generated search result screen (S42).

[0222] In this way, when the user sets both the side effect term and the drug name on the search word setting screen 43 of Figure 11, the terminal device 30 switches the display from the search word setting screen 44 to the final search result screen (see Figures 14-16) without displaying a screen like the drug name list screen 46 of Figure 13.This has the advantage that if the name of the drug you have looked up in advance is known, you can smoothly access the final search result screen.

[0223] Even in the case of this third flowchart, if the data group to be set (searched) is big data 5 storing medical-related literature information, the contents of the search result screen finally displayed on the terminal device 30 will be a literature search result screen 47 shown in FIG. 14, which can be used for screening work regarding side effects, etc. Also, if the data group to be searched is a hospital's electronic medical record DB, an electronic medical record search result screen 48 as shown in FIG. 16 will be displayed as the final search result screen at step S42, which can be used as a screen for notifying patients who are subject to an alert regarding side effects. Furthermore, if the target data group is a pharmacy's electronic medical history DB, an electronic medical history search result screen 49 as shown in FIG. 16 will be displayed as the final search result screen at step S42, which can also be used as an alert screen for notifying patients of side effects.

[0224] When multiple drug names are set (input) on the search word setting screen 43 (see FIG. 11), the processing shown in the third flowchart is performed in parallel for each drug name. Furthermore, when at least one drug name is set (input) on the search word setting screen 43 (see FIG. 11) and multiple side effect terms are input, the processing of the third flowchart for each drug name becomes the processing content when multiple side effect terms are input as described above.

[0225] Furthermore, when the automatic search mode is set at S14 in the first flowchart of Fig. 17, the subsequent processing will proceed according to the procedure shown in the second flowchart of Fig. 18. That is, in the automatic search mode of this embodiment, processing is performed using all of the side effect terms included in the drug link table 13 of Fig. 4 as search words (side effect terms for search), so that the processing is essentially the same as when multiple side effect terms are set as search words as described above, and the server device 10 performs the processing shown in S31-S35 in the second flowchart of Fig. 18.

[0226] However, in the automatic search mode, at step S36, drug names are not identified by an AND search-like process of multiple side effect terms found in the search process at step S35, but rather, for each side effect term found in the search process at step S35, the corresponding drug name is identified from the drug link table 13 in Figure 4. Then, at step S37, the server device 10 performs output processing to generate screen data for a drug name list screen 46 including a correspondence list 20 that associates the side effect terms found in the search process at step S35 with the drug names identified at step S36, and transmits the screen data to the terminal device 30.

[0227] However, in this automatic search mode, as described above, processing is performed on all side effect terms contained in the drug link table 13, so the number of drug names identified in step S36 is much greater in the automatic search mode than when side effect terms are entered in keyword search mode. Therefore, the number of drug names included in the correspondence list 20 in the screen data of the drug name list screen 46 generated in step S37 is also much greater than in keyword search mode, and the server device 10 transmits such screen data to the terminal device 30.

[0228] Furthermore, when the terminal device 30 in the automatic search mode accepts the selection of the execute button 45b on the automatic search confirmation screen 45 at step S6 of the first flowchart in FIG. 17, it proceeds to step S21 of the second flowchart in FIG. 18 without inputting a search word, and thereafter the processing proceeds according to this second flowchart.

[0229] However, the drug name list screen 46 of Figure 13 displayed on the display at step S22 contains a much larger number of rows in the correspondence list 20 arranged on that screen than the correspondence list 20 on the drug name list screen 46 displayed in the keyword search mode, as described above. Accordingly, the drug name list screen 46 displayed on the terminal device 30 in the automatic search mode contains a large number of drug names and a large number of side effect terms corresponding to each of those drug names, so that the user can check these large number of drug names and side effect terms by scrolling vertically.

[0230] Other processes relating to the second flowchart of the server device 10 and the terminal device 30 in the automatic search mode are the same as those in the keyword search mode described above. In this way, in the automatic search mode, the process proceeds automatically without the user having to input any keywords, and the drug name list screen 46 in Fig. 13 can be displayed, which is suitable for finding out which of all side effect terms included in the drug link table 13 correspond to the data group that is the setting target (search target), or for finding out the drug name corresponding to the matched side effect term.

[0231] Note that the present invention according to the first embodiment is not limited to the above-described contents, and various modified examples described below can be applied. For example, in the above description, various screens are displayed on the terminal device 30 based on screen data (for example, screen data corresponding to a web page of a website) provided by the server device 10. However, it is also possible to store the screen data that forms the basis of the various screens on the terminal device 30 side (for example, it is conceivable to store such screen data in the storage device 30g or to include it in the information search program 32), and to send only the information to be included (arranged) on each screen from the server device 10 to the terminal device 30.

[0232] In addition, in the above description, various processes that were performed by the server device 10 may be performed by the terminal device 30, with the server device 10 remaining in the role of a data server that stores the tables 12 to 16 and the databases 17, etc., and the terminal device 30 accessing the cloud system 4 (see Figure 1) and performing processing (search processing, information extraction processing, etc.) on a data group (for example, big data 5).

[0233] Furthermore, in the above explanation, when only side effect terms are input in keyword search mode, at step S37 of the second flowchart in Figure 18, the server device 10 performs output processing to generate screen data for a drug name list screen 46 including the correspondence list table 20 and send it to the terminal device 30 (S37), and at step S21 the terminal device 30 receives the screen data (S21: YES) and displays the drug name list screen 46 using the received screen data.However, as with the case when a drug name is input as a search word, the drug name list screen 46 may not be displayed, and the search result screen (see literature search result screen 47 in Figure 14, electronic medical record search result screen 48 in Figure 15, electronic medical history search result screen 49 in Figure 16, etc.) which is the final processing result may be displayed.

[0234] In this specification, in the second flowchart of Fig. 18, the drug names corresponding to the side effect terms found in the search process of S35 are identified from the drug link table 13 of Fig. 4 in step S36, but after completing the identification process of step S36, the processing of the server device 10 next jumps to step S39 rather than proceeding to step S37. That is, in step S39, the server device 10 extracts (detects) information (information related to medical literature if the data group to be searched is big data 5) containing both the side effect terms found in the search process of S35 and the drug names identified in step S36, generates screen data (search result screen data) corresponding to a search result screen (e.g., literature search result screen 47 shown in Fig. 14) that displays the extracted information (e.g., information related to medical literature) as a list, and transmits the generated screen data to the terminal device 30 (S40).

[0235] In this modified example, the processing of the terminal device 30 is as follows: after displaying the search word setting screen 43 shown in Fig. 11 at step S7 of the first flowchart in Fig. 17, a side effect term is input as a search word (S8: YES), and the input search word is transmitted to the server device 10, and the processing jumps to step S25 of the second flowchart in Fig. 18. Then, at step S25, it is determined whether or not screen data has been received from the server device 10, and if received (S25: YES), a search result screen (see Figs. 14-16) is generated using the received screen data and displayed on the display (S26).

[0236] Therefore, in this modified example, the display transition of the terminal device 30 proceeds from the search word setting screen 44 in Figure 12 to the search result screen shown in Figures 14-16, and the drug name list screen 46 in Figure 13 is not displayed, making it ideal when it is not necessary to present the drug name or when it is desired to smoothly present only the final search result screen to the user.

[0237] Note that this modification example of omitting the display of the drug name list screen 46 of Fig. 13 can also be applied to the automatic search mode described above. When this modification example is applied to the automatic search mode, the server device 10 identifies the corresponding drug name from the drug link table 13 of Fig. 4 for each side effect term found in the search process at step S35 in the second flowchart of Fig. 18 (S36), but the processing of the server device 10 does not proceed to step S37 but jumps to step S39.

[0238] At this stage S39, the server device 10 extracts (detects) information (information on medical-related literature if the search target is big data 5) that lists both the side effect terms found in the search process of S35 and the drug names identified at the stage S36, generates screen data corresponding to a search result screen (see Figures 14-16) that shows the extracted information as a list, and transmits this to the terminal device 30 (S40). However, since the screen data generated at this stage S40 reflects processing of all side effect terms included in the drug link table 13, the amount of information included in the list (information extracted at the stage S39) is significantly greater than the screen data generated at the stage S40 in the modified example in which only side effect terms are input in the keyword search mode described above.

[0239] On the other hand, in the processing of the terminal device 30 in this modified automatic search mode, if the selection operation of the Execute button 45b on the automatic search confirmation screen 45 is accepted at step S6 of the first flowchart in Fig. 17 (S6: YES), the processing jumps to step S25 of the second flowchart in Fig. 18. Then, at step S25, the terminal device 30 determines whether screen data has been received from the server device 10, and if so (S25: YES), generates a search result screen (see Figs. 14-16) using the received screen data and displays it on the display (S26). Note that, as described above, the amount of information included in the search result screen displayed at step S26 is far greater than that displayed in the modified case in which the display of the drug name list screen 46 is omitted in the keyword search mode.

[0240] In such a modified example in which the display of the drug name list screen 46 is omitted in the automatic search mode, if the user selects the execute button 45b on the automatic search confirmation screen 45 of Fig. 12(b) on the terminal device 30, the final search result screen is then displayed, thereby further enhancing the degree of automation in the automatic search mode. Note that the various modified examples of the first embodiment described above can also be used in combination as appropriate if they can be combined.

[0241] Furthermore, in the content based on the 1-3 flowchart in Figures 17-19 described above, the search type setting screen 44 in Figure 12(a) is displayed at step S3 in the first flowchart in Figure 17, allowing the user to select "keyword search" or "automatic search." However, when the data group to be searched is literature, etc., only "keyword search" may be available, as described above. In this case, when only "keyword search" is available, the processing on the terminal device 30 side in the first flowchart in Figure 17 skips steps S3-6, and when YES is selected at step S2 (when button 42a is selected), the processing proceeds to displaying the search word setting screen 43 at step S7. Then, when the back button 43c is selected on this search word setting screen 43 (S9: YES), the processing returns to the initial step S1.

[0242] Furthermore, when only this "keyword search" is performed, the processing on the server device 10 side omits the processing at steps S12-15 in the first flowchart, and without notification of the search type from the terminal device 30, processing starts from step S16, and the server device 10 enters automatic search mode upon the start of processing, and performs processing from step S17 onwards.

[0243] Furthermore, depending on the specifications of the information retrieval system 1, it is also conceivable that only "automatic search" will be performed. In this case where only "automatic search" is performed, the processing on the terminal device 30 side in the first flowchart of Fig. 17 omits the processing at stages S3, S4, and S7-10, and when YES is obtained at stage S2 (when button 42a is selected), the processing proceeds to display of the automatic search confirmation screen 45 at stage S5. Then, when the execute button 45b is selected on this search word setting screen 43 (S6: execute), the processing proceeds to the second flowchart of Fig. 18 without sending a notification of automatic search to the server device 10, and when the back button 45a is selected (S6: return), the processing returns to the initial stage S1.

[0244] Furthermore, when only this "automatic search" is performed, the processing in the first flowchart on the server device 10 side is performed without notification of the search type from the terminal device 30, and only the processing in step S14 is performed upon starting the processing, entering the automatic search mode, and the processing proceeds to the second flowchart in Figure 18, with the processing in the other steps in the first flowchart being omitted.

[0245] Furthermore, in the processing steps S33 and S34 of the server device 10 in the second flowchart of FIG. 18, a group of synonymous words showing synonymous relationships is identified based on the synonymous word table 16 of FIG. 7 (S33), one word is extracted for each identified group of synonymous words, and the extracted words are combined in the order of combination shown in the side effect word table 15 of FIG. 6 to generate "side effect terms with different spellings" by combining them so as to have different spellings from the "notified side effect terms (side effect terms for search)" and the "synonymous side effect terms" identified in step S31 (S34). However, the side effect terms generated are not limited to those with different spellings, and side effect terms may be generated using all possible combinations.

[0246] That is, in the above explanation, as shown in the example in Figures 8(a)-(c), the side effect term for search (side effect term notified to server device 10) consisting of a combination of multiple words is "leukopenia," and its synonymous side effect term (synonymous word combination side effect term) is "WBC decrease," and "leukopenia" is a combination of two words, "leukocyte" and "decrease," and "WBC decrease" is a combination of two words, "WBC" and "decrease." Of these four words, "leukocyte" and "WBC" form the first synonymous word group, and "decrease" and "decrease" form the second synonymous word group. Therefore, when one word is extracted from each of the first synonymous word group and the second synonymous word group and combined to generate a side effect term, the generated side effect term is written differently as "leukopenia" and "WBC decrease," but this is not limited to this, and all combinable combinations may be combined.

[0247] In the example shown in Figure 8(a)-(c), the first synonymous word group includes the two words "white blood cell" and "WBC," and the second synonymous word group includes the two words "decrease" and "decrease," so a total of four side effect terms can be generated by all possible combinations (two words x two words). These four words that can be generated are specifically "leukopenia," "decreased white blood cells," "decreased WBC," and "decreased WBC."

[0248] Then, for the side effect terms generated from all combinable combinations as described above, it is conceivable to perform a search process using all the generated side effect terms as search words in the search word search process at step S35 of flowchart 2. That is, in the above explanation, the search words at step S35 were "side effect terms notified to server device 10 (side effect terms for search)," "synonymous side effect terms (synonymous word combination side effect terms)" identified at step S31, and "side effect terms with different spellings" identified at step S34, but since the side effect terms generated from all combinable combinations as described above are equivalent to these, the search process will be the same even if all the generated side effect terms are used as search words.

[0249] By generating side effect terms in all combinable combinations in this way, it is not necessary to generate side effect terms that are spelled differently from the "side effect terms for search" and "synonymous side effect terms," ​​thereby reducing the processing burden on the server device 10 in the process of generating side effect terms with different spellings. Furthermore, if all generated side effect terms are used in the search process at step S35, side effect terms with different spellings will naturally be included. If all generated side effect terms are used in the search process at step S35 without using the "side effect terms for search" notified by the terminal device 30 or the "synonymous side effect terms" identified at step S31, the generation of side effect terms at step S34 can be smoothly connected to the search process at step S35, which is also favorable for the internal processing of the server device 10.

[0250] In addition, in the above explanation, the "side effect term for search" is a side effect term formed by combining multiple words so that search processing can be performed that includes "side effect terms with different spellings" in the search words. However, at step S31 of the second flowchart in Figure 18, when identifying synonymous side effect terms using the synonymous side effect table 14 in Figure 5, if a side effect term formed by combining multiple words is identified, a side effect term consisting of a single word can also be used as the "side effect term for search."

[0251] For example, assume that when performing a "keyword search," the user sets a side effect term consisting of a single word, "TEN," as the "side effect term for search" on the search word setting screen 43 of FIG.

[0252] For this side effect term "TEN," three synonymous side effect terms can be identified: "toxic epidermal necrolysis," "toxic epidermal necrolysis," and "Lyell's syndrome" based on the synonymous side effect table in Figure 5. As shown in side effect word table 15 in Figure 6, the first, "toxic epidermal necrolysis," is a combination of four words: "toxicity," "epidermal," "necrosis," and "syndrome," the second, "toxic epidermal necrolysis," is a combination of five words: "toxicity," "epidermal," "necrosis," "lysis," and "syndrome," and the third, "Lyell's syndrome," is a combination of four words: "Lyell" and "syndrome." Therefore, these three synonymous side effect terms are combinations of multiple words (corresponding to word-combined side effect terms).

[0253] Therefore, since "TEN," which is the "side effect term for search," is composed of a single word, it is not possible to generate a side effect term (e.g., a side effect term with a different spelling) by identifying a synonymous word group for each of these three synonymous word-combined side effect terms and extracting and combining one word from the synonymous word group. However, since each of these three synonymous word-combined side effect terms is composed of multiple words combined, it is possible to generate a side effect term (e.g., a side effect term with a different spelling) using these. In this way, when generating side effect terms from multiple synonymous word-combined side effect terms identified based on the synonymous side effect table, it is preferable to apply the method described in the above-mentioned modified example (a method of generating side effect terms by combining multiple words constituting each word-combined side effect term in all combinable combinations in the combination order shown in the side effect word table 15) in order to improve the efficiency of the generation process.

[0254] To explain the specific details of generating side effect terms that will be search words (search words used in the search process at step S35) using these three synonymous word combination side effect terms, when three word combination side effect terms are used, side effect terms will be generated using a total of three generation patterns.

[0255] First, in the first generation pattern, regarding the first "toxic epidermal necrolysis" and the second "toxic epidermal necrolysis", both words share the same beginning "toxic epidermal necrolysis" up to the word "necrosis", and both share the same ending "shou", so all possible combinations of the words that make up these two words result in only two words, "toxic epidermal necrolysis" and "toxic epidermal necrolysis", and this first generation pattern cannot generate side effect terms with different spellings from these two words.

[0256] As for the second generation pattern, for the first "toxic epidermal necrolysis" and the third "Lyell's syndrome", by combining each word in all possible combinations in the order of combinations shown in side effect word table 15, a total of four side effect terms can be generated: "toxic epidermal necrolysis", "toxic epidermal necrolysis syndrome", "Lyell's syndrome", and "Lyell's disease". Of these four words, "toxic epidermal necrolysis syndrome" and "Lyell's disease" are side effect terms with different spellings than "toxic epidermal necrolysis" and "Lyell's syndrome". In this generation pattern, the first term, "toxic epidermal necrolysis," is identified by decomposing it into two words, "toxic epidermal necrolysis" and "syndrome," and the second term, "Lyell's syndrome," is identified by decomposing it into two words, "Lyell" and "syndrome." For these four identified words, the first group of synonyms is "toxic epidermal necrolysis" and "Lyell," and the second group of synonyms is "syndrome" and "syndrome," and the side effect term generation process is performed.

[0257] As for the third generation pattern, for the second, "toxic epidermal necrolysis" and the third, "Lyell's syndrome," by combining each word in all possible combinations in the order of combinations shown in side effect word table 15, a total of four side effect terms can be generated: "toxic epidermal necrolysis," "toxic epidermal necrolysis syndrome," "Lyell's syndrome," and "Lyell's disease." Of these four words, "toxic epidermal necrolysis syndrome" and "Lyell's disease" are side effect terms with different spellings than "toxic epidermal necrolysis" and "Lyell's syndrome." In this generation pattern, the first term, "toxic epidermal necrolysis," is identified by decomposing it into two words, "toxic epidermal necrolysis" and "syndrome," and the second term, "Lyell's syndrome," is identified by decomposing it into two words, "Lyell" and "syndrome." For these four identified words, the first group of synonyms is "toxic epidermal necrolysis" and "Lyell," and the second group of synonyms is "syndrome" and "syndrome," and the side effect term generation process is performed.

[0258] When the side effect terms generated as described above are organized so as to avoid duplication, a total of six side effect terms, namely "toxic epidermal necrolysis," "toxic epidermal necrolysis syndrome," "toxic epidermal necrolysis," "toxic epidermal necrolysis syndrome," "Lyell's syndrome," and "Lyell's disease," can be generated from the three synonymous word combination side effect terms, "toxic epidermal necrolysis," "toxic epidermal necrolysis syndrome," "Lyell's syndrome," and "Lyell's disease." Of these six side effect terms, "toxic epidermal necrolysis syndrome," "toxic epidermal necrolysis syndrome," and "Lyell's disease" are side effect terms with different spellings than the first three synonymous word combination side effect terms.

[0259] In this way, when "TEN" is set by the user as a side effect term for a search on the search word setting screen 43, the processing described above is performed at steps S31-34 of the second flowchart in Figure 18, and at step S35, a total of seven side effect terms, including the six generated side effect terms and "TEN," are used as search words to perform the search processing.

[0260] Furthermore, for the above-mentioned "side effect terms for search," when identifying synonymous side effect terms using the synonymous side effect table 14 of Figure 5, the modified process of identifying side effect terms formed by combining multiple words and generating side effect terms can also be applied when side effect terms consisting of multiple words are used as "side effect terms for search."

[0261] For example, assume that when performing a "keyword search," the user sets the side effect term "toxic epidermal necrolysis," which is made up of multiple words, as the "side effect term for search" on the search word setting screen 43 in Figure 11.

[0262] For this side effect term "toxic epidermal necrolysis," three synonymous side effect terms can be identified: "toxic epidermal necrolysis," "Lyell's syndrome," and "TEN" based on the synonymous side effect table 14 in Figure 5. Of these three synonymous side effect terms, "toxic epidermal necrolysis" and "Lyell's syndrome" are side effect terms consisting of multiple words, and therefore can be used to generate side effect terms. Furthermore, "toxic epidermal necrolysis" set by the user is also a side effect term consisting of multiple words, and therefore can be used to generate side effect terms. Note that "TEN" identified based on the synonymous side effect table 14 is a side effect term consisting of a single word, and therefore is not used to generate side effect terms.

[0263] Therefore, in this case too, side effect terms will be generated using a total of three generation patterns. In the first generation pattern, two side effect terms will be used: "toxic epidermal necrolysis" set by the user and "toxic epidermal necrolysis" identified based on synonymous side effect table 14. In the second generation pattern, two side effect terms will be used: "toxic epidermal necrolysis" set by the user and "Lyell's syndrome" identified based on synonymous side effect table 14. In the third generation pattern, "toxic epidermal necrolysis" and "Lyell's syndrome," both identified based on synonymous side effect table 14, will be used.

[0264] Of these three generation patterns, the first and second generation patterns use "search side effect terms" set by the user and synonymous word combination side effect terms identified from the synonymous side effect table 14, and therefore correspond to the side effect term generation pattern described in the above-mentioned embodiment.On the other hand, the third generation pattern uses word combination side effect terms identified based on the synonymous side effect table 14, and therefore corresponds to the side effect term generation pattern described in the above-mentioned modified example.Therefore, this example is a mixture of the two generation patterns.

[0265] Furthermore, even in this example, when the first generation pattern of "toxic epidermal necrolysis" and "toxic epidermal necrolysis" is used, only "toxic epidermal necrolysis" and "toxic epidermal necrolysis" can be generated, and no alternative spellings of side effect terms can be generated with this first generation pattern. Furthermore, when the second generation pattern of "toxic epidermal necrolysis" and "Lyell's syndrome" is used, a total of four side effect terms can be generated: "toxic epidermal necrolysis," "toxic epidermal necrolysis syndrome," "Lyell's syndrome," and "Lyell's disease." Of these four words, "toxic epidermal necrolysis syndrome" and "Lyell's disease" are alternative spellings of side effect terms to "toxic epidermal necrolysis" and "Lyell's syndrome."

[0266] Furthermore, when the third generation pattern of "toxic epidermal necrolysis" and "Lyell's syndrome" is used, a total of four side effect terms can be generated: "toxic epidermal necrolysis," "toxic epidermal necrolysis syndrome," "Lyell's syndrome," and "Lyell's disease." Of these four terms, "toxic epidermal necrolysis syndrome" and "Lyell's disease" are side effect terms with different spellings than "toxic epidermal necrolysis" and "Lyell's syndrome." In this example, if the side effect terms generated in this way are organized so that there are no overlaps, the six side effect terms generated, "toxic epidermal necrolysis," "toxic epidermal necrolysis syndrome," "toxic epidermal necrolysis," "toxic epidermal necrolysis syndrome," "Lyell's syndrome," and "Lyell's disease," plus the single word "TEN," make up a total of seven side effect terms, and the search process at step S35 of the second flowchart in Figure 18 is performed.

[0267] Furthermore, when performing a search process by including the generated multiple side effect terms in search words at step S35 of the second flowchart in Figure 18, it is preferable to perform the search process by using these multiple side effect terms as search words in an order according to the level values ​​of the words contained in each side effect term (the importance level values ​​shown in the side effect word table 15 in Figure 6), in order to improve search accuracy and efficiency.

[0268] Specifically, as in the modified example described above, when a user performs a "keyword search" and sets "TEN" as the "side effect term for search" on the search word setting screen 43 of Figure 11, the process goes through steps S31 to S34 of the second flowchart in Figure 18, and at step S35, the search process is performed using the seven side effect terms "TEN," "toxic epidermal necrolysis," "toxic epidermal necrolysis syndrome," "toxic epidermal necrolysis syndrome," "Lyell's syndrome," and "Lyell's disease" as search words.

[0269] As shown in side effect word table 15 in Figure 6, among these seven side effect terms, "TEN" has a level value of "1," and the level values ​​of the constituent words of "toxic epidermal necrolysis" are "1," "2," "3," and "4," so its level value is the sum of these values, "10 (1 + 2 + 3 + 4)." Similarly, the level value of "toxic epidermal necrolysis syndrome" (the sum of the level values ​​assigned to each word; the same applies below) is "8," "toxic epidermal necrolysis" has a level value of "15," "toxic epidermal necrolysis syndrome" has a level value of "12," "Lyell's syndrome" has a level value of "3," and "Lyell's disease" has a level value of "5."

[0270] Since the smaller the level value of each word shown in the side effect word table 15 in Figure 6, the higher the importance, the smaller the level value (total value) of the seven side effect terms mentioned above is determined to be the more important. Therefore, when these seven side effect terms are arranged from lowest level value to highest importance, the order is as follows: first, "TEN (level value 1)," second, "Lyell's syndrome (level value 3)," third, "Lyell's disease (level value 5)," fourth, "toxic epidermal necrolysis syndrome (level value 8)," fifth, "toxic epidermal necrolysis (level value 10)," sixth, "toxic epidermal necrolysis syndrome (level value 12)," and seventh, "toxic epidermal necrolysis (level value 15)."

[0271] Therefore, at step S35 in the second flowchart of Figure 18, when performing search processing using a total of seven side effect terms, the search processing is performed using each side effect term as a search word in the order described above.Therefore, even when multiple side effect terms are used as search words, the search processing can be performed in order of importance, which results in efficient search processing and helps to obtain accurate search results.

[0272] Furthermore, the tables 13-16 shown in Figures 4-7 used in the present invention are not limited to the table configurations described above, and other configurations may be applied as modified examples. For example, the synonymous side effect table 14 in Figure 5 is not limited to a configuration in which multiple synonymous side effect terms are shown in a single line, and as another configuration example, a side effect term and its synonymous side effect term may be shown in one-to-one correspondence.

[0273] Fig. 20 shows a partial example of a modified synonymous side effect table 54, which is an example of a configuration showing synonymous side effect terms in such one-to-one correspondence. In summary, the synonymous side effect table 14 shown in Fig. 5 shows synonymous side effect terms in a horizontal array, while the synonymous side effect table 54 in Fig. 20 has a table configuration in which synonymous side effect terms are shown in a vertical array in each row.

[0274] Specifically, in the synonymous side effect table 54 of Figure 20, side effect terms with associated side effect codes and their synonymous side effect terms (which also have associated side effect codes) are configured in a one-to-one correspondence in the same row. For example, the side effect term "toxic epidermal necrolysis (side effect code 000006)" is associated one-to-one with its synonymous side effect term "toxic epidermal necrolysis (side effect code 000007)" in the same row, and in the next row, "toxic epidermal necrolysis (side effect code 000006)" is associated one-to-one with its synonym "Lyell's syndrome" and "toxic epidermal necrolysis (side effect code 000008)," and in the next row, "toxic epidermal necrolysis (side effect code 000006)" is associated one-to-one with its synonym "TEN" (side effect code 000009). In this way, the synonymous side effect table 54 of the modified example has a one-to-one correspondence between synonymous side effect terms in the same row, which has the advantage that synonymous side effect terms can be grasped for each row.

[0275] Furthermore, in addition to each of the tables 13-16 used in the present invention being an individual, independent table, it is also possible to combine (merge) the contents of multiple related tables into a single table, and in such a combined table, it is also preferable to introduce appropriate codes to indicate the relationship between the information within the tables.

[0276] 21 shows a partial example of a merged table 19 (synonymous side effect word table) configured by combining the contents of the synonymous side effect table 14 of FIG. 5 and the contents of the side effect word table 16 of FIG. 6. This merged table 19 has, from left to right, a side effect code column 19a, a side effect synonym code column 19b, a side effect term column 19c, and a word column 19d. The side effect code column 19a stores the side effect code of each side effect term as information for identifying the side effect term. The side effect synonym code column 19b stores the side effect synonym code for identifying synonymous side effect terms. The side effect term column 19c stores individual side effect terms. Similar to the side effect word table 16 of FIG. 6, the word column 19d indicates the individual words (words with attached importance levels) used to construct the side effect terms and the order in which these words are combined.

[0277] In the merge table 19 configured as above, the sections corresponding to the side effect code column 19a, side effect synonym code column 19b, and side effect term column 19c correspond to the contents of the synonymous side effect table 14, and the remaining sections excluding the side effect synonym code column 19b correspond to the contents of the side effect word table 16. Therefore, the merge table 19 is configured such that the side effect synonym code and side effect term correspond to the side effect code in the same row, and also corresponds to each word (words with an importance level attached) that makes up the side effect term in the same row while also indicating the order of combination, and as a whole, it is the side effect word table 16 in FIG. 6 to which the side effect synonym code column 19b has been added.

[0278] As a specific example of the contents of merge table 19, the side effect code "000001" is associated in the same row with the side effect term "leukopenia" which has a side effect synonym code of "F000001", and in the word column 19d, the side effect term "leukopenia" is associated with two words, "white blood cell (level value is 1)" and "decreased (level value is 2)", which indicates that "leukopenia" is the combination of the words "white blood cell" and "decreased" in that order.

[0279] In addition, in the same line, the side effect code "000002" is associated with the side effect term "WBC decrease" with the side effect synonym code "F000001", and in the word column 19d, it is associated with two words, "WBC (level value is 1)" and "decrease (level value is 2)".

[0280] The side effect synonym code for "leukopenia" above is "F000001," and the side effect synonym code for "decreased WBC" is also "F000001." Therefore, by being assigned the same side effect synonym code, "leukopenia" and "decreased WBC" are indicated as synonymous side effect terms. This is also true for other side effect terms. "Low red blood cell count" and "decreased RBC" are both indicated as synonymous side effect terms, since their side effect synonym code is "F000002." Furthermore, the four side effect terms "toxic epidermal necrolysis," "toxic epidermal necrolysis," "Lyell's syndrome," and "TEN" are all indicated as synonyms, since their side effect synonym code is "F000006." Merge Table 19 thus has the advantage of being able to show synonymous side effect terms in a single table, as well as the order in which each word in a side effect term is combined.

[0281] As described above, there are various variants of the present invention according to the first embodiment, and of course, these various variants can be applied in combination as appropriate, provided that they do not conflict with each other in terms of content. By combining and using various variants in this way, it is possible to further develop the various variants and provide specifications that meet the needs of users. [Example]

[0282] 22 is a schematic diagram showing an overall system configuration including an example of an information retrieval system 100 according to a second embodiment (Example 2) of the present invention. The information retrieval system 100 according to the second embodiment is basically based on the configuration according to the first embodiment described above, but is characterized in that it also uses symptom terms (including terms for initial symptoms) as search words. In the following description of the second embodiment, the same reference numerals as in the first embodiment will be used for the same components as in the first embodiment.

[0283] The information retrieval system 100 according to the second embodiment is composed of a server device 110 (side effect information retrieval device) that provides search results and the like, and a user terminal device 30, and performs processing on a data group that is the subject of the search processing. As in the first embodiment, the data group that is the subject of the processing is stored in a cloud (e.g., a cloud system 4) or the like (e.g., big data 5). However, in the second embodiment, such a data group includes information that describes symptom terms (including initial symptom terms; the same applies below) that describe symptoms (including initial symptoms) in addition to drug names and side effect terms, and in the second embodiment, symptom terms are also used as search words.

[0284] The basic hardware configuration of the server device 110 according to the second embodiment is the same as that of the server device 10 according to the first embodiment shown in Fig. 3, and includes an MPU 110a, a communication module (communication module 10b), a RAM (RAM 10c), an input interface (input interface 10e), an output interface (output interface 10f), and a storage unit 110g. However, in the second embodiment, the number of tables included in the table DB 111 stored in the storage unit 110g is greater than that in the first embodiment. Note that, in the second embodiment, the combination of the server device 110 and the table DB 111 functions as the side effect information retrieval system 109 according to the second embodiment, and the server device 110 and the table DB 111 may have separate hardware configurations, as in the first embodiment.

[0285] The tables included in the table DB 111 include the drug link table 13, synonymous side effect table 14, side effect word table 15, and synonym word table 16, which are also used in the first embodiment, and in addition there are tables for the second embodiment, such as a side effect symptom link table 113, a symptom table 114, a symptom synonym word table 115, and a symptom word pronunciation table 116.

[0286] In the second embodiment, the user database 17 and the screen database 118 for the second embodiment are also stored in the storage unit 110g as databases, and the screen database 118 for the second embodiment stores screen data corresponding to the processing content of the second embodiment. Also, the storage unit 110g stores, as programs, a server OS program P and a search program 112 for the second embodiment (corresponding to the computer program according to the present invention), and the search program 112 for the second embodiment defines processing corresponding to the second embodiment, and the MPU 110a functions as various means based on the defined processing content.

[0287] 23 shows an example of part of the contents of the side effect symptom link table 113 stored in the table DB 111 in the storage unit 110g. The side effect symptom link table 113 shows symptom terms corresponding to various side effect terms, storing each term in correspondence with the other, and also associates a symptom code and a symptom term corresponding to the symptom code with a side effect code and a side effect term corresponding to the side effect code (various corresponding codes and terms are stored in the same horizontal row). The symptom terms shown in this side effect symptom link table 113 are included in the symptom table 114, which will be described later.

[0288] As an example, as shown in FIG. 23, the side effect symptom link table 113 stores the symptom term "feeling tired (symptom code S001)" in association with the side effect term "leukopenia (side effect code 000001)" which is made up of a combination of multiple words, stores the symptom term "feeling chills (symptom code S002)" which is made up of a combination of multiple words in association with the same side effect term "leukopenia (side effect code 000001)" as above, and further stores the symptom term "feeling fever (symptom code S003)" which is made up of a combination of multiple words in association with the same side effect term "leukopenia (side effect code 000001)" as above.

[0289] As shown in the example of Figure 23, in addition to cases where multiple symptom terms are associated with one side effect term (e.g., "leukopenia"), there are also cases where only one symptom term is associated with one side effect term. By performing such associations, the side effect symptom link table 113 is able to indicate a side effect term corresponding to the symptom term. Note that while the example of Figure 23 only indicates one side effect term, "leukopenia," an actual side effect symptom link table 113 stores multiple types of side effect terms, and associates each side effect term with a related symptom term. Also, the example of the side effect symptom link table 113 in Figure 23 shows three symptom terms associated with the side effect term "leukopenia," but in reality, many more symptom terms are associated.

[0290] 24 shows an example of part of the contents of symptom table 114 stored in table DB 111 in storage unit 110g. Symptom table 114 includes various symptom terms, and for symptom terms formed by combining multiple words, the individual words used in the combination are displayed, along with the importance of each word in the symptom term and the order of combination.

[0291] That is, for each symptom term, the symptom table 114 lists each word that constitutes that symptom term on the same line, indicating that each of these words is used in combination. Furthermore, the symptom table 114 assigns a two-level level to each word that constitutes a symptom term, indicating the degree to which it best represents the meaning of that symptom term, as a measure of the importance of each word in the symptom term. These levels are prefixed with an S to distinguish them from the levels in the side effect word table 15 described above. The word that best represents the meaning of the symptom term is assigned the number "S1," while the other words are assigned the number "0."

[0292] For example, for the symptom term "tired" with the symptom code "S001", the symptom table 114 places the word "tired" in the same row in association with it, and since the symptom term "tired" is composed of only one word, only the word "tired" is placed in association with it. Also, since there is only one word, the level of the word "tired" is "S1", indicating that the word "tired" is the word that best expresses the meaning of the symptom term.

[0293] As another specific example, for the symptom term "feeling chills" with symptom code "S002," symptom table 114 places the words "chills," "is," and "do" in the same row in association with each other, indicating that these three words are used in conjunction. Symptom table 114 assigns a level of "S1" to the word "chills," indicating that it is the word that best represents the meaning of the symptom term "feeling chills," while assigning a level of "0" to the other words "is" and "do," indicating that they are semantically unimportant.

[0294] As yet another specific example, for the symptom term "running a fever" with symptom code "S003," symptom table 114 places the words "fever," "ga," and "depart" in the same row in association with each other, indicating that these three words are used in conjunction. Additionally, symptom table 114 assigns a level of "S1" to the word "fever," indicating that it is the word that best represents the meaning of the symptom term "running a fever," while assigning a level of "0" to the other words "ga" and "depart," indicating that they are semantically unimportant.

[0295] Furthermore, in the symptom table 114, the words arranged in the same row in association with the symptom term indicate the order of combination from left to right. For example, for the symptom term "feeling chills" with the symptom code "S002" described above, the words "chills," "is," and "do" are arranged in that order from left to right in the same row. Therefore, the symptom table 114 also indicates that in the symptom term to which these words are combined, the word "chills" is the first word, the second word is "is," and the third word is "do."

[0296] 25 shows an example of part of the contents of the symptom synonym table 115 stored in the table DB 111 in the storage unit 110g. The symptom synonym table 115 shows words that are synonymous, and for example, shows words that are synonymous (in this case, equivalent to synonymous words) with words that best represent the meaning of the symptom term among the words included in the symptom table 114 described above, that is, words with a level of S1 (in this case, equivalent to symptom words) in the symptom synonym table 115. More specifically, the symptom synonym table 115 shows that each word included in the same row (same line) is synonymous, and each word is assigned a code (symptom word code) that identifies the word.

[0297] For example, the symptom synonym word table 115 indicates that the word "tired" is synonymous with the word "fatigue" that is included in and associated with it on the same line, indicating that "tired" has a symptom word code of "S0010," and that "fatigue" has a symptom word code of "S0011." The symptom synonym word table 115 also indicates that the word "chills" is synonymous with the word "chills" that is included in and associated with it on the same line, indicating that "chills" has a symptom word code of "S0020," and that "chills" has a symptom word code of "S0021."

[0298] Note that symptom synonym word table 115 includes the word "fever" with symptom word code "S0030," but since no synonymous words can be expected for the word "fever," no other words that can be associated with the word "fever" are placed in the same row as the word "fever." In this way, symptom synonym word table 115 also includes words for which no synonyms can be expected, and in this case, symptom synonym word table 115 indicates that no synonyms exist.

[0299] 26 shows an example of part of the contents of the symptom word pronunciation table 116 stored in the table DB 111 in the storage unit 110g. The symptom word pronunciation table 116 indicates the words (corresponding to pronunciation words) that best represent the meaning of the symptom term among the words included in the symptom table 114 described above, i.e., words assigned a level of S1 (words including symptom words and words that are synonymous with them (synonymous words)). In this embodiment, words that represent pronunciations in hiragana are used. More specifically, in the symptom word pronunciation table 116, words that represent the pronunciation of words that represent the meaning of symptoms are included in the same row as the words that represent the meaning of the symptoms, and a symptom word code is assigned to the word that represents the pronunciation.

[0300] For example, the symptom word reading table 116 indicates that the word expressing the symptom "tired" is read using the hiragana word "tired" that is included in the same row. In this case, since the word expressing the symptom "tired" is originally written in hiragana, the reading word is also associated with the same hiragana word "tired." Note that the symptom word reading table 116 also assigns an identification code (reading code) to the reading word, and indicates the reading code "S0010A" for the word expressing the symptom "tired," which is the word symptom code with "A" added to the end.

[0301] As another specific example, the symptom word reading table 115 indicates that the reading of the word "fatigue" is represented by the hiragana word "kentai" included in association with the same line, and indicates that "fatigue" has a symptom word code of "S0011" and that "kentai" has a reading identification code of "S0011A." As yet another specific example, the symptom word reading table 115 indicates that the reading of the word "chills" is represented by the hiragana word "samuke" included in association with the same line, and indicates that "chills" has a symptom word code of "S0020" and that "samuke" has a reading identification code of "S0020A."

[0302] Next, each process defined by the search program 112 stored in the storage unit 110g will be described. The search program 112 according to the second embodiment uses some of the processes defined by the search program 12 according to the first embodiment (see FIG. 1 ) and further includes additional processes related to symptom terms. The unique processes according to the second embodiment include a search process related to symptom terms and a process for identifying side effect terms based on the results of the search process. Furthermore, among the processes defined by the search processing program 12 according to the first embodiment, those also used in the second embodiment include a user authentication process, a search process related to side effect terms, a process for identifying drug names based on the search results, and a process for extracting information from a data set based on the search results. The MPU 110a of the server device 110 functions as various means by executing each of these processes according to the search program 112.

[0303] In the processing of the second embodiment, user authentication processing and processing for setting the search type are also performed based on the specifications of the search program 112. That is, either a "keyword search" or an "automatic search" is performed depending on the type of data group to be searched (whether the search target is literature, etc.), the system specifications, etc., or either a "keyword search" or an "automatic search" is set by the user. When a "keyword search" is performed, the server device 110 proceeds with processing in keyword search mode, and when an "automatic search" is performed, the server device 110 proceeds with processing in automatic search mode.

[0304] 27 shows a search word setting screen 144 of the second embodiment that is displayed on the access source terminal when processing proceeds in keyword search mode. In the second embodiment as well, when the server device 110 proceeds with processing in keyword search mode, screen data corresponding to the search word setting screen 144 is sent to the access source terminal, causing the search word setting screen 144 to be displayed on the access source terminal. The search word setting screen 144 of this second embodiment differs from the search word setting screen 43 of the first embodiment shown in FIG. 11 in the input content of the first input field, which is a symptom term input field 144a (note that the second input field is a drug name input field 144b, as in the first embodiment).

[0305] At least one symptom term can be entered into this symptom term input field 144a, and when a symptom term is entered and the search button 144d is selected, the entered symptom term is notified (transmitted) from the access source terminal to the server device 110 as a search word.

[0306] When the server device 110 (MPU 110a) receives notification of a search word from the access source terminal, if a symptom term (corresponding to a symptom term for search) in the received search word is a combination of multiple words, the MPU 110a, as a combined word identification means, performs a process of breaking down such symptom term into the individual words used in the combination and identifying them based on the symptom table 114 shown in Figure 24.

[0307] For example, if the received search word is the symptom term "feeling chills" (symptom term for search), the MPU 110a refers to the symptom table 114 and performs a process of identifying three words, "chills," "but," and "do," which are arranged in association with the same row as "feeling chills." If the received search word is the symptom term "running a fever" (symptom term for search), the MPU 110a performs a process of identifying three words, "fever," "but," and "come out," from the symptom table 114. If the received search word is the symptom term "tired" (symptom term for search), the MPU 110a determines that the symptom term "tired" is simply the word "tired" in this identification process, because only the word "tired" is arranged in association with the same row as "feeling chills" in the symptom table 114.

[0308] Next, the MPU 110a performs a process of specifying synonymous words (synonymous words) for a word (symptom word) that best represents the meaning of the symptom related to the symptom term for search, from among the words specified from the symptom term for search, using the symptom synonymous word table 115 shown in Fig. 25. In detail, after the symptom term for search is broken down and specified into individual words using the symptom table 114 in Fig. 24, the MPU 110a specifies words assigned a level of "S1" (corresponding to symptom words that best represent the meaning of the symptom) from among the specified words, and performs a process of specifying "synonymous words (synonymous words)" for the specified word (symptom word) from the symptom synonymous word table 115 in Fig. 25.

[0309] Specifically, when the symptom term used for search is "feeling chills," as described above, the MPU 110a breaks it down into three words, "chills," "is," and "does," and identifies them. In the symptom table 114 of FIG. 24, the word "chills" is identified as the "word that best expresses the meaning of the symptom (symptom word)" based on the additional information of level "S1." In the symptom synonym word table 115 of FIG. 25, "chills," which is placed in the same row as the word "chills," is identified as the "synonymous word (synonymous word)."

[0310] In addition, when the symptom term used for search is "having a fever," as described above, it is broken down into three words, "fever," "ga," and "get," and from among these, MPU 110a identifies the word "fever" as the "word that best expresses the meaning of the symptom (symptom word)" because it is at the "S1" level in symptom table 114. However, since there is no word associated with the word "fever" in the same row in symptom synonym word table 115, it is determined that there are no synonyms for the word "fever."

[0311] Furthermore, as described above, the symptom term used for the search, "tired," is determined to be a single word, "tired," and since the word "tired" is at the "S1" level, MPU 1110a identifies it as "the word that best expresses the meaning of the symptom (symptom word)," and identifies "fatigue," which is placed in the same row as the word "tired" from the symptom synonym word table 115, as "a synonymous word (synonymous word)."

[0312] Then, the MPU 110a performs a process of identifying, from among the words identified from the symptom term for search, a "word that best indicates the meaning of the symptom (symptom word)" for that symptom term for search and a "word that is a synonym (synonym word)" for that word, based on the symptom word pronunciation table 116 of Figure 26.

[0313] Specifically, when the symptom term used for search is "feeling chills," the processing of the MPU 110a described above targets "chills," which is the "word that best indicates the meaning of the symptom (symptom word at level S1)," and "chills," which is the "synonymous word (synonym word)" of "chills." From the symptom word pronunciation table 116 of FIG. 26, the MPU 110a identifies, for the word "chills," the "word representing the pronunciation (pronunciation word)" "samu-ke," which is arranged in the same row, and for the word "chills," it identifies the "word representing the pronunciation (pronunciation word)" "okan," which is arranged in the same row.

[0314] In addition, when the symptom term used for search is "having a fever," the processing of MPU 110a described above identifies only "fever," which is the "word that best indicates the meaning of the symptom (symptom word at level S1)," and determines that there are no synonyms. Therefore, MPU 110a identifies, from symptom word pronunciation table 116, the "word that indicates the pronunciation (pronunciation word)" "netsu," which is arranged in the same row as the word "fever."

[0315] Furthermore, if the symptom term used for search is "tired," the processing by MPU 110a described above targets the word "tired," which is one of the "words that best represent the meaning of the symptom (symptom words at level S1)," and "fatigue," which is one of the "synonyms (synonymous words)" of "tired." From the symptom word pronunciation table 116, MPU 110a identifies the "word representing the pronunciation (pronunciation word)" of the hiragana "tired" that is associated with the same line for the symptom word "tired," and identifies the "word representing the pronunciation (pronunciation word)" of "kentai" that is associated with the same line for the synonym word "fatigue." In this example, for the symptom word "tired" at level "S1", a word with the same pronunciation (reading word) as "tired" is identified, so MPU 110a ultimately identifies the reading word as the same spelling for the symptom word "tired" at level S1, and does not identify the word "tired" which has the same spelling twice.

[0316] Next, the MPU 110a sets the "words that best represent the meaning of the symptom (symptom words)," "synonymous words (synonymous words)," and "words that represent pronunciation (pronunciation words)" identified in the above process as search words, and performs a search process as word search means to determine whether such search words are included in the data group to be searched. Note that the process of setting the search target is the same as in the first embodiment, as described above.

[0317] This search process determines whether the data group to be searched (e.g., Big Data 5) contains any of the search words "words that best indicate the meaning of symptoms (symptom words)," "synonymous words (synonymous words)," and "words that indicate pronunciation (pronunciation words)" (whether the search process matches).

[0318] Specifically, if the symptom term to be searched is "feeling chills," then, after the above-described processing, the MPU 110a performs a search process on the data group using four words: "chills" as the "word that best describes the meaning of the symptom (symptom word)," "chills" as a synonym (synonym word), and "cold" and "okan" as "words that describe pronunciation (pronunciation word)." Also, if the symptom term to be searched is "having a fever," then the MPU 110a performs a search process using two words: "fever" as the "word that best describes the meaning of the symptom (symptom word)" and "fever" as a "word that describes pronunciation (pronunciation word)" (there are no matching "synonyms"). Furthermore, if the symptom term used for the search is "tired," MPU 110a will perform the search process using three words: "tired" as the "word that best indicates the meaning of the symptom (symptom word)," "fatigue" as a synonym (synonym word), and "kentai" corresponding to "fatigue" as a word that indicates the pronunciation (pronunciation word).

[0319] If a matching word (a word that is found to be included in a data group) is found in the search process by the word search means of the MPU 110a, the MPU 110a, as symptom identification means, performs a process of identifying a symptom term in which the matching word is used, based on the symptom table 114 shown in Fig. 24. Then, the MPU 110a, as side effect identification means, performs a process of identifying a side effect term corresponding to the symptom term identified in the process by the symptom identification means of the MPU 110a, based on the side effect symptom link table 113 shown in Fig. 23.

[0320] Specifically, when a search process is performed on a data group using four words, "chills" as the "word that best indicates the meaning of the symptom (symptom word)," "chills" as a synonym (synonym word), and "cold" and "okan" as "words that indicate pronunciation (pronunciation word)," if "chills" matches (hits), the MPU 110a refers to the symptom table 114 in Figure 24 to identify the symptom term "feeling chills" that is associated with and located in the same row as the word "chills." The MPU 110a then identifies a side effect term corresponding to the identified symptom term "feeling chills" based on the side effect symptom link table 113 in Figure 23. In this case, the side effect term "leukopenia" that is associated with and located in the same row as the symptom term "feeling chills" is identified.

[0321] Furthermore, when searching a data group using the above-mentioned four words "chills," "chills," "cold," and "mother," if "chills" is found, the MPU 110a will refer to the symptom table 114 in Fig. 24, but the word "chills" is not included in this symptom table 114. If the hit word is not included in the symptom table 114, the MPU 110a next determines whether the hit word "chills" is included in the symptom synonym word table 115 in Fig. 25.

[0322] As shown in Figure 25, the symptom synonym word table 115 includes the word "chills," so the MPU 110a identifies the synonymous word "chills" that is associated with the word "chills" in the same row. After identifying the synonymous word "chills," the MPU 110a again refers to the symptom table 114 in Figure 24 and identifies the symptom term "feeling chills" that is associated with the word "chills" in the same row. From this point on, as in the process described above, the side effect term "leukopenia" is identified based on the side effect symptom link table 113 in Figure 23.

[0323] Furthermore, when searching a data group using the four words mentioned above, "chills," "chills," "cold," and "mother," if "cold" is found, the MPU 110a refers to the symptom table 114 and determines that the word "cold" is not included. Then, it determines whether the hit word "cold" is included in the symptom synonym word table 115 of FIG. 25.

[0324] As shown in Figure 25, the symptom synonym word table 115 does not include the word "cold." In this case, the MPU 110a further determines whether the hit word "cold" is included in the symptom word pronunciation table 116 of Figure 26. As shown in Figure 26, the symptom word pronunciation table 116 includes the word "cold." Therefore, the MPU 110a identifies the word "chills" associated with "cold" in the same row, based on the symptom word pronunciation table 116. Thereafter, the MPU 110a performs the same process as when "chills" was found to be a match, thereby identifying the symptom term "feeling chills." Finally, the side effect term "leukopenia" associated with the symptom term "feeling chills" is identified based on the side effect symptom link table 113 of Figure 23.

[0325] Furthermore, when searching a data group using the four words mentioned above, "chills," "chills," "cold," and "mother," if "mother" is found, the MPU 110a will refer to the symptom table 114 and determine that the word "mother" is not included. Then, it will determine whether the hit word "mother" is included in the symptom synonym word table 115 of FIG. 25.

[0326] As shown in Figure 25, "okan" is not included in the symptom synonym word table 115, so the MPU 110a next determines whether the word "okan" is included in the symptom word pronunciation table 116 of Figure 26. Then, as shown in Figure 26, the word "okan" is included in the symptom word pronunciation table 116, so the MPU 110a identifies the word "chills" which is associated with the word "okan" and placed in the same row. After this, the MPU 110a performs the same process as when "chills" was found to be a match, and identifies the symptom term "feeling chills," and finally identifies the side effect term "leukopenia" which is associated with the symptom term "feeling chills" based on the side effect symptom link table 113 of Figure 23.

[0327] When searching a data group using two words, "fever" as the "word that best indicates the meaning of the symptom (symptom word)" and "netsu" as the "word that indicates the pronunciation (pronunciation word)," if "fever" matches (hits), the MPU 110a refers to the symptom table 114 in Fig. 24 to identify the symptom term "fever" that is associated with and placed in the same row as the word "fever." Then, based on the side effect symptom link table 113 in Fig. 23, the MPU 110a identifies the side effect term "leukopenia" that is associated with and placed in the same row as the identified symptom term "fever."

[0328] Furthermore, when searching a data group using the two words "netsu" (heat) and "fetsu" (heat), if "fetsu" is found, the MPU 110a will refer to the symptom table 114 in Fig. 24. However, since the symptom table 114 does not contain the word "fetsu," the MPU 110a then determines whether or not "fetsu" is included in the symptom synonym word table 115 in Fig. 25. Then, since the symptom synonym word table 115 does not contain the word "fetsu," the MPU 110a further determines whether or not "fetsu" is included in the symptom word pronunciation table 116 in Fig. 26.

[0329] 26, since the symptom word pronunciation table 116 includes the word "fever," the MPU 110a identifies the word "heat" that is associated with "fever" in the same row, based on the symptom word pronunciation table 116. Thereafter, the MPU 110a performs the same process as when "fever" was found to be a match, to identify the symptom term "fever developing," and finally identifies the side effect term "leukopenia" that is associated with the symptom term "fever developing," based on the side effect symptom link table 113 in FIG.

[0330] Furthermore, when searching a data group using three words, "weary" as the "word that best indicates the meaning of the symptom (symptom word)," "fatigue" as a synonym (synonym word), and "kentai" as a word that indicates pronunciation (pronunciation word), if "weary" matches (hits), the MPU 110a refers to the symptom table 114 in Fig. 24 to identify the symptom term "weary" that is associated with and located in the same row as the word "weary." Then, based on the side effect symptom link table 113 in Fig. 23, the MPU 110a identifies the side effect term "leukopenia" that is associated with and located in the same row as the identified symptom term "weary."

[0331] Furthermore, when a search process for a data group is performed using the three words "weary," "fatigue," and "health" described above, if "fatigue" is found, the MPU 110a will refer to the symptom table 114 of FIG. 24. However, since the symptom table 114 does not include the word "fatigue," the MPU 110a next determines whether "fatigue" is included in the symptom synonym word table 115 of FIG. 25. Then, since the symptom synonym word table 115 includes the word "fatigue," the MPU 110a identifies the synonymous word "weary" associated with the word "fatigue" in the same row. Thereafter, the MPU 110a performs the same process as when "weary" was found described above to identify the symptom term "weary," and finally identifies the side effect term "leukopenia" associated with the symptom term "weary" based on the side effect symptom link table 113 of FIG. 23.

[0332] Furthermore, when searching a data group using the three words "tired," "fatigue," and "kentai" as described above, if "kentai" is found, the MPU 110a refers to the symptom table 114 in Fig. 24, but since the symptom table 114 does not contain the word "kentai," it next determines whether or not "kentai" is included in the symptom synonym word table 115 in Fig. 25. Then, since the symptom synonym word table 115 does not contain the word "kentai," the MPU 110a further determines whether or not "kentai" is included in the symptom word pronunciation table 116 in Fig. 26.

[0333] 26, since the symptom word pronunciation table 116 includes the word "kentai," the MPU 110a identifies the word "tai" which is associated with "kentai" in the same row, based on the symptom word pronunciation table 116. Thereafter, the MPU 110a performs the same process as when "tai" was found to be the case, to identify the symptom term "lazy," and finally identifies the side effect term "leukopenia" which is associated with the symptom term "lazy," based on the side effect symptom link table 113 in FIG.

[0334] When searching for a data group using multiple words, such as when searching for a data group using the four words described above, when searching for a data group using three words, or when searching for a data group using two words, it is possible that two or more words will be matched in the search results. In this case, when two or more words are matched, the above-described process is performed for each of the two or more matched words, until the side effect term is finally identified. In addition, when identifying side effect terms when two or more words are matched as described above, it is possible that different side effect terms will be identified.

[0335] Furthermore, when a search process for each of the above-mentioned words is performed on a data group, it is possible that the search result will be no match (no hits). In this case, since the process cannot proceed to the identification of a side effect term, the process ends when no match is found, and the server device 110 (MPU 110a) generates screen data corresponding to a screen indicating no match and transmits it to the access source terminal.

[0336] Once the side effect terms are identified through the above-described process, the process is essentially the same as that of the first embodiment (when only symptom terms are entered in keyword search mode). That is, the server device 110 (MPU 110a) uses the side effect terms identified through the above-described process as side effect terms for search, performs the same process as that of the first embodiment using the side effect terms for search, identifies drug names, generates a drug name list screen (see FIG. 13), and transmits screen data of the generated drug name list screen to the access source terminal (the server device 110 performs the process corresponding to S31 to S37 of the second flowchart in FIG. 18). Note that, if multiple different side effect terms are identified through the above-described process, each of the different side effect terms is used as a side effect term for search, and the same process as that when multiple side effect terms are entered in keyword search mode in the first embodiment is performed on these side effect terms for search.

[0337] In addition, the same processing as in the first embodiment is performed on the accessing terminal, displaying a drug name list screen based on the sent screen data, and sending a notification of the drug name selected by the user on that drug name list screen to the server device 110.

[0338] Then, as in the first embodiment, the server device 110 performs a process of extracting information (e.g., medical-related literature) from the data group that includes both the drug name sent from the accessing terminal and the side effect terms identified in the above process, and generates screen data for a search result screen that includes a list that displays the extracted information, etc. (see Figure 14).

[0339] Furthermore, in the second embodiment, when a search process for a data group is performed using each of the above-mentioned words, if any of the words matches (hits), the information (e.g., medical-related literature) contained in the data group that describes the matching word is identified, and if the identified information is included in the list on the above-mentioned search result screen, the MPU 110a generates screen data for the search result screen so that the matching word and symptom terms related to that word are also included in the list in association with that information.

[0340] Fig. 28 shows a literature search result screen 147, which is an example of a search result screen in the second embodiment. This literature search result screen 147 has a configuration basically similar to that of the literature search result screen 47 shown in Fig. 14, and is used when a collection of medical-related literature A, B, C, etc. describing side effect terms (for example, big data 5 shown in Fig. 2(a)) is set as a data group to be searched, and the screen content corresponds to when a word related to a symptom term used in the search process (such as fever, fever) is described in medical-related literature A.

[0341] As in the first embodiment, the literature information list 121 included in this literature search result screen 147 is provided with a word column 121a indicating search words, etc., a literature name column 121b indicating the name of the literature, and a bibliographic information column 121c indicating bibliographic information about the literature, and the word column 121a associated with medical-related literature A arranged in the literature name column 121b contains (stores) a symptom term "fever" and a synonym of the symptom term "fever" in addition to the side effect term "leukopenia." Note that, as in the literature search result screen 47 of Fig. 14, this literature search result screen 147 also has a search setting symptom column 147e above the literature information list 121, which displays the symptom term of the search word entered by the user on the search word setting screen 144 of Fig. 27 (Fig. 28 shows an example in which the symptom term "fever" is entered by the user).

[0342] It is also assumed that the symptom term "feeling chills" is entered on the search word setting screen 144, and in such a case, the symptom term "feeling chills" will be displayed in the search setting symptom field 147e of the literature search result screen 147, and in this case, in addition to the side effect term "leukopenia," the symptom term "feeling chills" and "chills," which is a synonym of the word "chills" included in the symptom term, will be arranged (stored) in the word field 121a. Therefore, when the user looks at this literature search result screen 147, he or she can see that medical-related literature A has content related to the symptom term "feeling chills" and that the word "chills" is used.

[0343] Furthermore, the above content has been explained based on the case where, in keyword search mode, the user inputs one search word (symptom term for search) into the symptom term input field 144a in the search word setting screen 144 of Figure 27 (for example, when one symptom term such as "having a fever" or "having chills" is input), but it is of course also possible that multiple symptom terms are input as search words.

[0344] When multiple symptom terms are entered as search words in this way, the process for identifying the side effect terms described above is performed for each symptom term, and then the same process as in the first embodiment is performed from the side effect terms identified in each process to identify the drug name, and a drug name list screen (see Figure 13) that summarizes the drug names corresponding to each process is generated, and the screen data of the generated drug name list screen is transmitted to the access source terminal. Note that even when multiple symptom terms are entered as search words, after the screen data of this drug name list screen is transmitted, the same process as described above is performed, and finally a search result screen (for example, a screen equivalent to the literature search result screen 147 shown in Figure 28) is presented to the access source terminal.

[0345] 27, when a symptom term is input into the symptom term input field 144a and a drug name is input into the drug name input field 144b, the MPU 110a performs processing as a search term receiving means, and performs processing similar to that when a drug name is input into the search word setting screen 44 of the first embodiment. That is, when a side effect term is identified through the above-mentioned processing, the MPU 110a identifies a drug name corresponding to the identified side effect term based on the drug link table 13 (see FIG. 4), but does not generate a drug name list screen at a subsequent stage, and performs processing as drug name identification means to identify, from the identified drug names, a drug name that is the same as the drug name input by the user (the drug name for search).

[0346] Then, MPU 110a performs the above-mentioned processing to extract information (e.g., pharmaceutical-related literature information) containing both the side effect term and the identified name of the same drug from the data group, generates screen data corresponding to the search result screen (e.g., literature search result screen 147), and transmits (outputs) it to the access source terminal (output processing).

[0347] In addition, when multiple drug names are entered in the drug name input field 144b of the search word setting screen 144, the same processing as in the first embodiment is performed, and the MPU 110a extracts information containing the identified side effect term and the multiple entered drug names from the data group, generates screen data corresponding to the search result screen, and sends it to the accessing terminal.

[0348] Furthermore, the above description is for the keyword search mode, but depending on the type of data group to be searched and the system specifications, processing will proceed in automatic search mode, and the processing content of server device 110 in automatic search mode will be as follows.

[0349] In the automatic search mode, as in the first embodiment, the server device 110 first transmits screen data corresponding to the automatic search confirmation screen (see FIG. 12(b)) to the access source terminal, and when it receives a notification of "execute" from the access source terminal, the server device 110 proceeds with the execution of the automatic search process. Note that the screen data corresponding to the automatic search confirmation screen is transmitted to the access source terminal in this manner in order to confirm with the user whether or not to actually execute an automatic search, since the processing in the automatic search mode places a heavy processing load on the server device 110. In the automatic search process of the second embodiment, the symptom terms included in the symptom table 114 in FIG. 24 (or the side effect symptom link table 113 in FIG. 23) are used as search words (symptom terms for search) for processing.

[0350] In one example of the second embodiment, all symptom terms included in the symptom table 114 are used as search words. Therefore, if the symptom table 114 includes 1000 symptom terms, the MPU 110a will perform the process described in the keyword search mode above (the process from inputting one search word to identifying a side effect term) for each of the 1000 symptom terms.

[0351] Furthermore, a characteristic process in the automatic search mode of the second embodiment is a process of assigning a score to the results of a search process of a data group for each word, based on the number of hits for each word. As a specific example, if a search process is performed on a data group (e.g., big data 5 including a large number of medical-related documents) using the word "chills," and the word "chills" is found in 38 medical-related documents among the large number of medical-related documents, the word "chills" will be matched (hit) a total of 38 times, and the MPU 110a will assign a score of 38 points to the word "chills." Similarly, if a search process is performed on the word "fatigue" and the word "fatigue" is found in 27 medical-related documents, the word "fatigue" will be matched (hit) a total of 27 times, and the MPU 110a will assign a score of 27 points to the word "fatigue."

[0352] Furthermore, in processing when a side effect term is identified, if there are multiple symptom terms associated with the identified side effect term in the side effect symptom link table 113, the MPU 110a performs processing to calculate the ratio between the number of identified symptom terms that use words found to be included in the search process and the number of symptom terms corresponding to the identified side effect term that are shown in the side effect symptom link table 113.

[0353] As a specific example, suppose the identified side effect term is "leukopenia," there are 10 symptom terms associated with "leukopenia" in the side effect symptom link table 113, and it is determined that the three words "tired," "chills," and "fever" are hits (matches) in the search process of the data group. In this example, the three symptom terms "tired," "chills," and "fever" are identified as symptom terms that use the three words "tired," "chills," and "fever" that were hit in the search process. Therefore, the MPU 110a calculates a ratio of 3 / 10 based on the number of identified symptom terms (3) and the number of symptom terms associated with "leukopenia" in the side effect symptom link table 113 (10). This calculated ratio can be an indicator of the suitability of the identified side effect term ("leukopenia" in this example). In the above case, if the only hit (match) in the data group search process is the word "fever," the symptom term that uses the word "fever" will be only one, "having a fever," and the ratio calculated by MPU 110a will be 1 / 10.

[0354] After identifying the side effect term, the process proceeds as in the first embodiment, where the MPU 110a identifies the drug name, generates a drug name list screen (see FIG. 13) that summarizes the drug names corresponding to each process, and transmits the screen data of the generated drug name list screen to the access source terminal. Then, based on the drug names sent from the access source terminal, the MPU 110a finally generates screen data corresponding to a search result screen (for example, a screen basically equivalent to the literature search result screen 147 shown in FIG. 28).

[0355] Fig. 28 shows a literature search result screen 147, which is an example of a search result screen according to the second embodiment. In this literature search result screen 147, words (words related to symptom terms) are arranged (stored) in a word column 121a in the literature information list 121. Although not shown in Fig. 28, when scores and ratios are calculated as described above, the MPU 110a generates screen data corresponding to the search result screen in which these scores and ratios are included in the word column 121a, and transmits (outputs) this data to the access source terminal.

[0356] For example, if the word "heat" is entered in the word column 121a and the score for the word "heat" is 38, the score will be entered after "heat" as "heat (38)." Therefore, by checking the scores entered after the words, the user can recognize the frequency of use of each word in the data set to be searched, which is useful for screening work.

[0357] In addition, if the symptom term "leukopenia" is entered as a specified side effect term in the word field 121a, and the percentage calculated from hits of the three words "tired," "chills," and "fever" is 3 / 10, the content will be "leukopenia (3 / 10)," with the percentage value entered after "leukopenia." Therefore, by checking the percentage value entered after the side effect term, the user can recognize the appropriateness of the specified side effect term based on an objective numerical value, which is useful for screening work.

[0358] Furthermore, the content of the second embodiment described above has been explained in the case where a data group to be searched is basically set as a data group containing literature information such as medical-related literature A, B, and C (for example, big data 5), ​​but as in the case of the first embodiment, the processing content will basically be the same even if a data group containing other different information is set as the search target.

[0359] 29 shows an electronic medical record search result screen 148 when the search target is a data group including patient information with patient names (for example, the X Hospital electronic medical record DB6 managed by the X Hospital's electronic medical record system in FIG. 2(b)). Like the electronic medical record search result screen 48 of the first embodiment (see FIG. 15), this electronic medical record search result screen 148 is a screen that arranges a patient information list 122 including a first column 122a indicating search words and the like used to extract the patient, a second column 122b indicating the patient's name and the like (a column indicating patient identification information), and a third column 122c indicating information regarding the patient's treatment and the like (medical information indicating hospital visit status, attending physician, etc.).

[0360] The electronic medical record search result screen 148 of the second embodiment is characterized by including in the first column 122a, words (words related to symptom terms) found in the search process and symptom terms related to those words. Therefore, the processing of the second embodiment is useful when searching for patients with symptoms related to symptom terms. (The electronic medical record search result screen 148 of the second embodiment also serves as a kind of alert screen that warns patients according to their symptoms.) The electronic medical record search result screen 148 has the same configuration as the electronic medical record search result screen 48 of the first embodiment shown in FIG. 15 except for the above-mentioned features. A scroll bar 148a, a back button 148b, a back button 148c, a view button 148d, and a search setting symptom column 148e (a column showing symptom terms for the search word entered by the user) are appropriately arranged on the screen.

[0361] 30 shows an electronic medical history search result screen 149, which, like the electronic medical history search result screen 49 of the first embodiment (see FIG. 16), is a search result screen when the search target is a data group containing a type of patient information different from that of the database of the electronic medical record system (for example, the Y Pharmacy electronic medical history DB7 managed by the Y Pharmacy's electronic medical history system in FIG. 2(c)). This electronic medical history search result screen 149 also has a patient information list 123 arranged thereon, which includes a first column 123a indicating the search words used to extract the patient, a second column 123b indicating the patient name, etc., and a third column 123c indicating information regarding the prescription of medicines, etc. (medical information indicating the name of the prescribing doctor, the name of the pharmacist, etc.).

[0362] 29, the electronic medical history search result screen 149 of the second embodiment is characterized in that, compared to the electronic medical history search result screen 49 of the first embodiment, the first column 123a includes words (words related to symptom terms) that were found in the search process and symptom terms related to those words. Therefore, even if a database managed by the electronic medical history system is set as the data group to be searched, the user can confirm the words (words related to symptom terms) that were found in the search process and symptom terms related to those words, making the processing of the second embodiment useful when searching for patients included in the database managed by the electronic medical history system from the perspective of symptoms (the electronic medical history search result screen 149 of the second embodiment also serves as a kind of alert screen that warns patients according to their symptoms). In addition, the electronic medical history search result screen 149 has the same configuration as the electronic medical record result screen 49 of the first embodiment in Figure 16, except for the above-mentioned parts, and a scroll bar 149a, a back one button 149b, a back to beginning button 149c, a view button 149d, and a search setting symptom column 149e (a column showing symptom terms of the search word entered by the user) are appropriately arranged on the screen.

[0363] Furthermore, the terminal device 30, which corresponds to the access source terminal used in the second embodiment, is basically the same as the terminal device 30 in the first embodiment, and performs processing to display a screen according to screen data sent from the server device 110, and also performs processing to accept various operations from the user based on the displayed screen and return information indicating the contents of the accepted operations to the server device 110, so details will be omitted.

[0364] Flowcharts 4-6 shown in Figures 31-33 show the main processing steps of the side effect information search method related to the second embodiment of the present invention, and represent the process in which the above-mentioned server device 110 and terminal device 30 process the data group (e.g., big data 5) to be searched for using symptom terms, ultimately leading to the display and output of a search result screen (see Figures 28-30, etc.).

[0365] The fourth flowchart in Fig. 31 corresponds to the first flowchart of the first embodiment shown in Fig. 17, and the basic processing flow is similar to that of the first flowchart, with each process from S101 to S110 on the terminal device side in the fourth flowchart corresponding to each process from S1 to S10 in the first flowchart. Note that a difference in content between the fourth flowchart and the first flowchart is, for example, that what is displayed at step S107 of the fourth flowchart is search word setting screen 144 shown in Fig. 27, and at least a symptom term is entered as a search word at step S107.

[0366] The processing on the server device 110 side in the fourth flowchart is also substantially the same as that in the first flowchart, and the processing in steps S112 to S118 in the fourth flowchart corresponds to the processing in steps S12 to S18 in the first flowchart. Note that one difference between the server side processing in the fourth flowchart and the first flowchart is that the search word in the notification received in step S117 includes a symptom term.

[0367] The fifth flowchart in Figure 32 corresponds to the second flowchart of the first embodiment shown in Figure 18, and the processing on the server device 110 side includes processing specific to the second embodiment in the first half (S131-S136), but the processing on the server device 110 side in the second half (S137-S138) and the terminal device 110 side is basically the same as that in the second flowchart.

[0368] As in the first embodiment, this fifth flowchart also illustrates the processing performed when a symptom term is set as a search word in keyword search mode without including a drug name. First, the server device 110 (similar to the MPU 110a; the same applies below) identifies the symptom term (symptom term for search) of the search word included in the received notification by breaking it down into individual words to be used for combining, based on the symptom table 114 in Fig. 24 (S131). Then, the server device 110 identifies, based on the symptom table 114, the word that best represents the meaning of the symptom related to the symptom term, and identifies synonymous words (synonymous words) for the identified word (symptom word) from the symptom synonymous word table 115 shown in Fig. 25 (S132).

[0369] Furthermore, the server device 110 identifies "words that represent pronunciations (pronunciation words)" for each of the "words that best represent the meaning of the symptom (symptom words)" and "synonymous words (synonymous words)" based on the symptom word pronunciation table 116 in Fig. 26 (S133). Then, the server device 110 performs a search process on the data group (e.g., big data 5) to be searched, determined at step S112 of the fourth flowchart, using the "words that best represent the meaning of the symptom (symptom words)," "synonymous words (synonymous words)," and "words that represent pronunciations (pronunciation words)" as search words (S134).

[0370] Then, the server device 110 identifies symptom terms related to the words found in the search process at step S134 based on the symptom table 114 shown in Fig. 24 (S135).The server device 110 then identifies side effect terms corresponding to the symptom terms identified at step S135 based on the side effect symptom link table 113 shown in Fig. 23 (S136).

[0371] From this step S136 onwards, the processing is equivalent to that of the second flowchart of Figure 18 as described above, and the side effect terms identified in step S136 are used as side effect terms for search, and the server device 110 performs processing from steps S31 to S37 of the second flowchart in step S137 (for example, in step S31 of the second flowchart in step S137, the side effect terms identified in step S136 are used as side effect terms for search, and synonymous side effect terms are identified based on the synonymous side effect table 14), and generates screen data corresponding to the drug name list screen (see Figure 13) and transmits it to the terminal device 30.

[0372] If the terminal device 30 accepts only symptom terms as search words at step S108 of the fourth flowchart in Figure 31 and sends a notification of the accepted symptom terms to the server device 110 at step S110, it proceeds to step S121 of the fifth flowchart in Figure 32, and at steps S121 to S124, it performs processing at steps S21 to S24 of the second flowchart and sends a notification of information such as the name of the drug selected by the user to the server device 110.

[0373] Furthermore, the server device 110 basically performs the processing of steps S38 to S40 of the second flowchart in Figure 18 from steps S138 to S140 after completing the processing of step S137. However, when generating screen data for a search result screen (for example, each search result screen 147-149 shown in Figures 28-30) at step S140, the fifth flowchart in the second embodiment is characterized in that the generated screen data also includes words (matched words) that were hit in the search processing of the data group at step S134, as well as symptom terms corresponding to the hit words.

[0374] The terminal device 30 basically performs the processing of steps S25 and S26 of the second flowchart in steps S125 and S126 after completing the processing of step S124, but the search result screen displayed in step S126 (for example, each search result screen 147-149 shown in Figures 28-30) is characterized in that it includes words (matched words) that were hit in the search processing of the data group in step S134, as well as symptom terms corresponding to the hit words.

[0375] 28, the search result screen displayed in this manner can be used to check the symptom terms entered by the user and documents (medical-related documents) that contain words related to those symptom terms from the document information list 121 included in the document search result screen 147, which is useful for screening various symptoms including initial symptoms. Note that, like in the first embodiment, the document search result screen 147 of the second embodiment also allows the contents of a document to be checked by selecting the name of the document included in the document information list 121.

[0376] Furthermore, when the data group to be searched is a hospital's electronic medical record database (for example, X Hospital Electronic Medical Record DB6 in FIG. 2(b)), the search result screen displayed at step S126 is the electronic medical record search result screen 148 shown in FIG. 29. This screen also allows the user to confirm the symptom terms entered by the user for the patient name, as well as words related to those symptom terms, and is therefore useful for searching for patients who may be experiencing various symptoms, including early symptoms. From the perspective of searching for patients in this way, the patient names included in the electronic medical record search result screen 148 represent patients who are the target of an alert for the symptom terms in the search word, and the electronic medical record search result screen 148 functions as an alert screen for patients according to the symptoms set by the user.

[0377] Furthermore, when a pharmacy's electronic medical history database (e.g., Y Pharmacy's electronic medical history DB7) is the data group to be searched, the search result screen displayed at step S126 is the electronic medical history search result screen 149 shown in FIG. 30. This screen also allows the user to confirm the symptom terms entered by the user for the patient name, as well as words related to those symptom terms, and is therefore useful for finding patients who may experience various symptoms, including initial symptoms. In this case, as with the electronic medical record search result screen 148 described above, from the perspective of searching for patients, the patient names included in the electronic medical history search result screen 149 represent patients who are the target of an alert for the symptom terms in the search word, and the electronic medical history search result screen 149 functions as an alert screen for patients according to the symptoms set by the user.

[0378] Depending on the symptom term entered at step S108 of the fourth flowchart in Figure 31, there may be cases where the search process at step S134 of the fifth flowchart in Figure 32 does not yield a hit (i.e., the search process does not match). If such a case occurs, at this stage, the server device 110 transmits screen data to the terminal device 30 indicating that the search process for the symptom term yielded no hits. Upon receiving this screen data, the terminal device 30 switches the display to a screen indicating that there were no search hits for the symptom term (a screen including a selectable confirmation button). Then, upon receiving a selection operation of the confirmation button on this screen, the process returns to step S107, and the search word setting screen 144 of Figure 27 is displayed again, providing the user with an opportunity to input another search word.

[0379] Furthermore, depending on the side effect term identified at step S136 of the fifth flowchart in Figure 32, there may be cases where the search process in step S137 (the search process corresponding to step S35) does not result in a hit (is not applicable in the search process). If such a case occurs, at this step, the server device 110 transmits screen data to the terminal device 30 indicating that there were no hits in the search process for the side effect term identified from the symptom term. Upon receiving this screen data, the terminal device 30 switches the display to a screen (including a selectable confirmation button) indicating that there were no hits for the side effect term identified from the symptom term. Then, upon receiving a selection operation of the confirmation button on this screen, the process returns to step S107, as described above, and the search word setting screen 144 is redisplayed, providing the user with an opportunity to input another search word.

[0380] Furthermore, depending on the drug name selected at step S123 in the fifth flowchart of Figure 32, there may be cases where the extraction (detection) process from the data group at step S139 does not find any relevant information. Even in such cases, the server device 110 transmits screen data to the terminal device 30 indicating that the drug name was not found, and upon receiving this screen data, the terminal device 30 switches the display to a screen indicating that there were no hits for the drug name (a screen including a selectable confirmation button). In this case, if the terminal device 30 receives a selection operation on the confirmation button on the screen indicating that there were no hits, the process returns to step S122, and the drug name list screen is displayed again, providing the user with an opportunity to select another drug name.

[0381] The sixth flowchart in Fig. 33 shows the processing in the keyword search mode when a drug name is input and set in addition to a symptom term on the search word setting screen 144 shown in Fig. 27 (processing when a drug name is included in the search word). In this sixth flowchart, the processing from S151 to S156 on the server device 110 side is the same as the processing from S131 to S136 in the fifth flowchart described above, and side effect terms are identified through processing based on the symptom terms for search input by the user (S156).

[0382] Then, in the step S157, the processing of steps S31 to S36 of the second flowchart of Figure 18 is performed, and in this case, in the step S35, the server device 110 performs a search process of the data group to be searched (e.g., big data 5) using side effect terms (e.g., "side effect terms for search," "synonymous side effect terms," ​​and "side effect terms with different spellings"), and further in the step S36, the name of the drug corresponding to the side effect term that is hit (matched) in the search process is identified from the drug link table 13 of Figure 4.

[0383] Then, the server device 110 identifies, from the identified drug names, a drug name that is the same (same spelling) as the drug name accepted upon receiving the notification (the drug name for search input and set by the user) (S158), and searches for and extracts (detects) information (e.g., information on medical-related literature) describing both the side effect term found in the search process at step S35 of step S157 and the drug name identified at step S158 from the data group to be searched (e.g., big data 5) (S159). Then, the server device 110 generates screen data (screen data of the search result) corresponding to a search result screen (e.g., literature search result screen 147 shown in FIG. 28) that shows the extracted (detected) information (e.g., information on medical-related literature) as a list, and transmits it to the terminal device 30 (S160).

[0384] On the other hand, when the terminal device 30 receives a symptom term and a drug name as search words at step S108 of the fourth flowchart in Fig. 31, it performs the process at step S110 and then proceeds to step S141 of the sixth flowchart in Fig. 33, where it determines whether or not it has received screen data for a search result screen sent from the server device 110. If it has not received the screen data (S141: NO), it enters a state of waiting for reception. If it has received the screen data for a search result screen (S141: YES), the terminal device 30 generates a search result screen (for example, the literature search result screen 147 in Fig. 28) based on the received screen data, and switches the display to show the generated search result screen (S142).

[0385] In this way, when both symptom terms and drug names are set on the search word setting screen 144, the drug name list screen is not displayed on the terminal device 30, and the display switches from the search word setting screen 144 to the final search result screen.Therefore, when the name of the drug you have looked up in advance is known, it is preferable to also enter the drug name in keyword search mode, as this will allow you to smoothly obtain the final search result screen.

[0386] Furthermore, when the automatic search mode is set at S115 in the fourth flowchart of Fig. 31, the subsequent processing steps proceed according to the procedure shown in the fifth flowchart of Fig. 32. That is, in the automatic search mode of the second embodiment, processing is performed using all of the symptom terms included in the symptom table 114 of Fig. 24 as symptom terms for search, so that the processing is essentially the same as when multiple symptom terms are set as search words, and the server device 110 repeatedly performs the processing shown in S131-S136 in the fifth flowchart of Fig. 32 the number of times equal to the number of symptom terms included in the symptom table 114.

[0387] However, as a process unique to the automatic search mode of the second embodiment, the server device 110 performs a search process for each word at the stage of S135, and when each word is hit (corresponding), the number of hits is assigned as a score (points) for that word (the word score is temporarily stored in a memory, etc.).

[0388] Furthermore, when a side effect term is identified in step S136, the server device 110 identifies, in step S137, the corresponding drug name for each identified side effect term from the drug link table 13 of Fig. 4, and performs output processing to generate screen data for a drug name list screen (see Fig. 13) including a correspondence list in which the side effect terms identified in step S136 and the drug names identified in step S137 are associated with each other, and transmit this screen data to the terminal device 30. As in the automatic search mode of the first embodiment, the number of drug names included in the screen data transmitted in this manner is far greater in the second embodiment than in the keyword search mode.

[0389] On the other hand, in the automatic search mode of the second embodiment, when the terminal device 30 receives a selection operation of the execute button for the automatic search at step S106 of the fourth flowchart in Fig. 31, it proceeds to step S121 of the fifth flowchart in Fig. 32, and thereafter the processing proceeds according to this fifth flowchart (the processing from steps S121 to S124 is performed). However, the drug name list screen displayed on the display at step S122 contains a much larger number of rows in the correspondence table arranged on that screen than in the keyword search mode, as described above (and accordingly the number of side effect terms contained in the drug name list screen is also much larger than in the keyword search mode).

[0390] Furthermore, after the processing of the terminal device 30 from S121 to S124 described above, the server device 110 performs the processing from S138 to S140. However, a feature of the automatic search mode of the second embodiment is that when generating screen data for a search result screen (for example, each search result screen 147-149 shown in Figures 28-30) at the S140 stage, the score (points) assigned to each word at the S135 stage described above is also included for the words to be included in the screen data.

[0391] The terminal device 30 then performs the processing in steps S125 and S126 to display the search result screen. The words related to symptom terms included in this search result screen are also shown with the above-mentioned score (points), so by checking this score (points), the user can understand the frequency with which words related to symptom terms appear. Even if the information included in the search result screen is more than in keyword search mode, referring to the score (points) can help the user to easily understand the content of the information.

[0392] In this way, in automatic search mode, processing proceeds automatically without the user having to enter any keywords, and a drug name list screen can be displayed on the terminal device 30. Therefore, this mode is suitable for those who want to know, for example, the words that make up all the symptom terms included in the symptom link table 114, that are listed in the information included in the data group (e.g., big data 5) that is the setting target (search target), or the drug names based on the search results.

[0393] The present invention according to the second embodiment is the same as the first embodiment except for the details described above. Furthermore, the present invention according to the second embodiment is not limited to the details described above, and various modified examples are applicable.

[0394] For example, it is possible to combine processing based on symptom terms in the second embodiment with processing based on side effect terms in the first embodiment. When combining the processing of the first and second embodiments in this way, it is possible to conceive of cases in which both processings are executed in parallel, cases in which the processing of the first embodiment is executed first and then the processing of the second embodiment, cases in which the processing of the second embodiment is executed first and then the processing of the first embodiment, etc. Regardless of how the first and second embodiments are combined, there will be parts of the processing that are common to both, and so it is intended to share the processing for such common parts.

[0395] Specifically, the processes at steps S1 to S18 in the first flowchart shown in Fig. 17 of the first embodiment are basically the same as the processes at steps S101 to S118 in the fourth flowchart shown in Fig. 31 of the second embodiment, so that one of them can be shared. For example, if the processes at steps S1 to S18 in the first flowchart of the first embodiment are performed, the results of that processing can also be used in the second embodiment, and the processes at steps S101 to S118 in the fourth flowchart can be omitted. However, in order to share the processes of the first and second embodiments, it becomes necessary to perform unique processing in accordance with the sharing.

[0396] Fig. 34 shows a search word setting screen 244 of a modified example, which is used for unique processing in accordance with sharing, and is configured as a combination of the search word setting screen 43 of the first embodiment shown in Fig. 11 and the search word setting screen 144 of the second embodiment shown in Fig. 27. When the keyword search mode is set in the modified example, such a search word setting screen 244 is displayed on the terminal device 30 at step S7 of the first flowchart (or S107 of the second flowchart).

[0397] This search word setting screen 244 includes an input field for side effect terms to search (first input field) 244a, an input field for drug names to search (second input field) 244b in the first embodiment, an input field for symptom terms to search (third input field) 244c, and an input field for drug names to search (fourth input field) 244d in the second embodiment, and also has a selectable back button 244e and search button 244f.

[0398] When a side effect term is input into the first input field 244a on the search word setting screen 244, the processing of the second flowchart shown in Fig. 18 is executed, and when a drug name is input into the second input field 244b in addition to inputting a side effect term into the first input field 244a, the processing of the third flowchart shown in Fig. 19 is executed. When a symptom term is input into the third input field 244c on the search word setting screen 244, the processing of the fifth flowchart shown in Fig. 32 is executed, and when a drug name is input into the fourth input field 244d in addition to inputting a symptom term into the third input field 244c, the processing of the sixth flowchart shown in Fig. 33 is executed.

[0399] In addition, if no side effect term is entered in the first input field 244a on the search word setting screen 244, the processing according to the first embodiment will not be executed, and if no symptom term is entered in the third input field 244c, the processing according to the second embodiment will not be executed.

[0400] Furthermore, when a side effect term is entered in the first input field 244a and a symptom term is entered in the third input field 244c on the search word setting screen 244, the processes shown in the second flowchart and the processes shown in the fifth flowchart will be executed as described above. However, the processes of S21-S26 and S37-S40 in the second flowchart and the processes of S121-S126 and S37-S140 in S137 in the fifth flowchart are basically the same, so these will also be shared by either one of them.

[0401] However, the screen data for the drug name list screen (e.g., drug name list screen 46 in FIG. 13) generated at S37 of the second flowchart (or S37 in S137 of the fifth flowchart) will contain two correspondence lists: a correspondence list according to the first embodiment including the drug names identified at S36 of the second flowchart, and a correspondence list according to the second embodiment including the drug names identified at S36 in S137 of the fifth flowchart. Accordingly, the drug name list screen displayed on the terminal device 30 at S22 of the second flowchart (or S122 of the fifth flowchart) will also contain two correspondence lists: a correspondence list according to the first embodiment and a correspondence list according to the second embodiment, allowing the user to check both the correspondence list according to the first embodiment and the correspondence list according to the second embodiment at the same time.

[0402] Therefore, notification of the drug names selected from the correspondence list according to the first embodiment and the correspondence list according to the second embodiment included in the displayed drug name list screen will be sent to the server device 110 at S24 of the second flowchart (or S124 of the fifth flowchart).

[0403] In addition, when the server device 110 receives notification of the drug names selected from the correspondence list according to the first embodiment and the correspondence list according to the second embodiment, at step S39 of the second flowchart (or step S139 of the fifth flowchart), information according to the first embodiment and information according to the second embodiment are extracted based on the received drug names, etc.

[0404] Then, at S40 of the second flowchart (or S140 of the fifth flowchart), the server device 110 generates screen data corresponding to a search result screen including two information lists: an information list (first information list) that is the processing result of the first embodiment including information according to the first embodiment, and an information list (second information list) that is the processing result of the second embodiment including information according to the second embodiment, and transmits the screen data to the terminal device 30.

[0405] As a result, at the stage S26 of the second flowchart (or S126 of the fifth flowchart), the terminal device 30 (or terminal device 30) displays a search result screen including two information lists: an information list of the processing results of the first embodiment (first information list) and an information list of the processing results of the second embodiment (second information list). This allows the user to check the processing results of each embodiment together, and efficiently perform screening tasks based on each search result and extract relevant patients.

[0406] Furthermore, on the search word setting screen 244 of Figure 34, when a side effect term is entered in the first input field 244a, a drug name is entered in the second input field 244b, a symptom term is entered in the third input field 244c, and a drug name is also entered in the fourth input field 244d, the processes shown in the third flowchart and the processes shown in the sixth flowchart will be executed as described above, but the processes of S41, S42, S57 to S59 of the third flowchart and the processes of S141, S142, S158 to S160 of the sixth flowchart are basically common, so these are also shared by either one.

[0407] However, even if a pharmaceutical name is entered on the search word setting screen 244, the server device 110 will extract (detect) information according to the first embodiment and extract (detect) information according to the second embodiment at S58 of the third flowchart (or S159 of the sixth flowchart), and will generate screen data according to a search result screen including two information lists: an information list (first information list) that is the processing result of the first embodiment including information according to the first embodiment, and an information list (second information list) that is the processing result of the second embodiment including information according to the second embodiment, and send this to the terminal device 30 (or terminal device 30).

[0408] As a result, at step S42 of the third flowchart (or step S142 of the sixth flowchart), the terminal device 30 displays a search result screen including two information lists: an information list of the processing results of the first embodiment (first information list) and an information list of the processing results of the second embodiment (second information list), and the processing results of each embodiment can be viewed together on a single search result screen.

[0409] In addition, in a modified example in which the processes of the first and second embodiments are combined, sharing of processes is possible even when the automatic search mode is set, and while sharing similar to that in the keyword search mode described above is performed, unique processing according to the modified example (such as displaying a pharmaceutical name list screen including two correspondence lists, one corresponding to the first embodiment and one corresponding to the second embodiment, and displaying a search result screen including two information lists, one information list of the processing results of the first embodiment (first information list) and one information list of the processing results of the second embodiment (second information list)) is performed.

[0410] 26 includes words whose pronunciations are written in hiragana, it is also possible to use words whose pronunciations are written in katakana instead of words whose pronunciations are written in hiragana, and it is also possible to include both words whose pronunciations are written in hiragana and words whose pronunciations are written in katakana in the word pronunciation table 116. Furthermore, in the content of the second embodiment, the various modified examples described in the first embodiment may also be applied to parts to which the various modified examples described in the first embodiment can be applied. [Example]

[0411] 35 is a schematic diagram showing the overall system configuration including an example of a side effect information retrieval system 200 according to a third embodiment (Example 3) of the present invention. The side effect information retrieval system 200 according to the third embodiment is basically based on the configuration according to the first or second embodiment described above, but is characterized in that it performs more strict processing for handling drug names.

[0412] That is, in the first and second embodiments, generic names, which are generally used names, are used for the names of pharmaceutical products. As an example of such generic names, the names of pharmaceutical products listed on the NHI drug price list are given, but generic names also include other ways of writing the names of pharmaceutical products listed on the NHI drug price list.

[0413] More specifically, several different types of names based on different viewpoints are generally used as common names for the same drug. In addition to the drug names listed in the drug price standard mentioned above, the different types of names include the notified names used when notified in the Official Gazette, the sales names (or trade names) given by the drug manufacturer (pharmaceutical company), and the prescription computer processing system drug names used in the prescription computer processing system (a system that processes medical fees using prescriptions, which are electronic medical fee statements).

[0414] These drug names listed in the NHI drug price list, notified drug names, brand names (trade names), and drug names in the electronic medical receipt processing system basically share the main characteristic parts of the name that identify the drug, but in many cases the additional parts other than the main parts are written differently. Note that, depending on the drug, there may be cases where two, three, or all of the drug names listed in the NHI drug price list, notified drug names, brand names, and drug names in the electronic medical receipt processing system are the same.

[0415] Therefore, if the notation of one of the names listed in the NHI drug price list, the notified name, the brand name, or the name of the medical receipt computer processing system (for example, notation based on the name of the medical receipt listed in the NHI drug price list) is used as the common name in the processing of the first or second embodiment described above, there is a risk of a search being missed if a different notation of the name (for example, notation based on the name of the medical receipt computer processing system) is used for the same drug in the data group to be searched, and therefore the content of the third embodiment (Example 3) is to perform processing that takes such search omissions into consideration. Note that in the following description of the third embodiment, the same symbols as in the first or second embodiment are basically used for parts that are similar to those in the first or second embodiment.

[0416] In order to prevent the above-mentioned search omissions, the server device 210 (side effect information retrieval device) included in the side effect information retrieval system 200 of the third embodiment has a basic hardware configuration equivalent to that of the server device 10 of the first embodiment (or the server device 110 of the second embodiment) shown in Fig. 3, but a common name link table 270 is newly stored in the table DB 211 stored in the storage unit 210g. Note that in the third embodiment, the combination of the server device 210 and the table DB 211 functions as the side effect information retrieval system 209 of the third embodiment, and the server device 210 and the table DB 211 may have separate hardware configurations, as in the first or second embodiment.

[0417] 35 is basically based on the server device 110 of the second embodiment, and includes an MPU 210a, a communication module (communication module 10b), a RAM (RAM 10c), an input interface (input interface 10e), an output interface (output interface 10f), and a storage unit 210g. A table DB 211 stored in the storage unit 210g stores tables 13-18 according to the first embodiment, and also stores tables 113-116 according to the second embodiment. The storage unit 210g also stores a search program 212 according to the third embodiment.

[0418] 36 shows an example of part of the contents of the common name link table 270, which is a feature of the third embodiment. The common name link table 270 arranges the names of the same drug, including the drug name listed in the drug price list, the notified name, the brand name, and the drug name used in the prescription computer processing system, in the same row, and associates them with each other.

[0419] In addition, the search program 212 of the third embodiment basically specifies the same processing content as the search program 12 of the first embodiment or the search program 112 of the second embodiment, but has the added feature that when processing related to pharmaceutical names, the MPU 210a uses the above-mentioned generic name link table 270 to identify other name notations.

[0420] For example, when a notification of a drug name (e.g., a name written based on the name of a drug listed in the drug price standard) entered in the drug name input field 43b of the search word setting screen 43 shown in Figure 11 of the first embodiment is received and transmitted from the terminal device 30 to the server device 210 (if step S18 in the first flowchart of Figure 17 is YES), the server device 210 (MPU 210a) will perform a process of searching whether the transmitted drug name is included in the generic name link table 270 based on the programming content specified by the search program 212.

[0421] If the name of the pharmaceutical product being processed matches (is hit) in the search process of this common name link table 270, the MPU 210a identifies other names (for example, names other than those listed in the drug price list) that are placed on the same line as the matched pharmaceutical product name in the common name link table 270, and performs a process to determine whether these identified name notations include any names that are different from the pharmaceutical product name transmitted from the terminal device 30.

[0422] If a different name is included, the MPU 210a performs processing by including the differently written drug name identified from the generic name link table 270 in the search words in addition to the drug name (the name of the drug listed in the drug price standard) transmitted from the terminal device 30 (for example, step S58 in the third flowchart of Figure 19).

[0423] By performing the processing according to the third embodiment, even if the same drug has multiple different spellings as generic names, there is an advantage in that it is possible to prevent search omissions in the search processing of a data group when the drug name is used as a search word. Note that if the different spellings of the drug name transmitted from the terminal device 30 are not included in the generic name link table 270, only the drug name transmitted from the terminal device 30 will be used as a search word.

[0424] As described above, the process of searching to see if the pharmaceutical name transmitted from the terminal device 30 is included in the generic name link table 270, and if the search returns a hit, identifying whether different name notations are included in the same row of the generic name link table 270, can also be applied when a pharmaceutical name (e.g., the name of a pharmaceutical listed in the drug price list) is entered in the pharmaceutical name input field 144b of the search word setting screen 144 shown in Figure 27 of the second embodiment (if step S118 in the fourth flowchart of Figure 31 is YES), or when a pharmaceutical name (e.g., the name of a pharmaceutical listed in the drug price list) is entered in the second input field 244b or fourth input field 244d of the search word setting screen 244 shown in Figure 34, which is a modified example of the second embodiment.

[0425] The processing according to the third embodiment described above can also be applied when the server device 210 generates screen data corresponding to the drug name list screen 46 shown in Fig. 13. That is, the drug name list screen 46 includes a correspondence list 20 that associates drug names with side effect terms, and the drug names (e.g., drug names listed in the NHI drug price list) shown in this correspondence list 20 are identified from the drug link table 13 in Fig. 4, for example, at step S36 of the second flowchart of the first embodiment shown in Fig. 18. The MPU 210a performs processing based on the programming content specified by the search program 212 to search for whether or not different spellings of the identified drug names (e.g., drug names listed in the NHI drug price list) are included in the generic name link table 270.

[0426] If names with different spellings are included, the MPU 210a will, for example, at step S37 of the second flowchart of the first embodiment, generate screen data for a drug name list screen including a correspondence table listing different spellings of drug names in addition to the drug names (e.g., drug names listed in the drug price list) identified from the drug link table 13 of Fig. 4, and transmit this screen data to the terminal device 30. The terminal device 30 will then receive this screen data and display the drug name list screen at step S22 of the second flowchart.

[0427] Compared to the drug name list screen 46 of the first embodiment in Figure 13, the drug name list screen displayed in the third embodiment displays multiple different spellings for the same drug, which may include the name of the spelling that the user normally uses, making it easier for the user to understand the contents of the drug and contributing to improved usability for the user.

[0428] Of course, the processing of the third embodiment regarding the above-mentioned pharmaceutical name list screen can also be applied to the second embodiment. For example, in the processing of step S36 included in step S137 of the fifth flowchart of the second embodiment shown in Figure 32, the MPU 210a searches whether the generic name link table 270 in Figure 36 contains a differently written pharmaceutical name for a pharmaceutical name identified from the pharmaceutical link table 13 (for example, the name of a pharmaceutical listed in the drug price list).

[0429] If the search results include names written in different ways, screen data for a drug name list screen including a correspondence table listing the drug names written in different ways in addition to the drug names (e.g., drug names listed in the drug price list) identified from the drug link table 13 is generated and transmitted to the terminal device 30. Based on this screen data, the terminal device 30 displays the drug name list screen at step S122 of the fifth flowchart. By applying the processing according to the third embodiment to the content of the second embodiment, multiple different ways of writing the same drug are listed on the drug name list screen, making it easier for the user to understand the contents of the drug.

[0430] The processing according to the third embodiment described above can also be applied when displaying a final search result screen. For example, in the first embodiment, if big data 5 including a large number of medical-related documents is set as a data group to be searched, a literature search result screen 47 shown in Fig. 14 is displayed as a final search result screen (one example), and by performing the processing according to the third embodiment on a drug name (for example, a drug name listed in the NHI drug price list) included in the word field 21a in the literature information list 21 on this literature search result screen 47, it becomes possible to display other names written in different ways.

[0431] Such a literature search result screen 47 searches and extracts information (information related to medical literature) that contains both the search words, the drug name and side effect terms, from the data group to be searched (for example, processing at step S39 of the second flowchart in Figure 18 or step S56 of the third flowchart in Figure 19).If information that contains both the drug name and side effect terms is found (a hit), the MPU 110a performs processing related to the third embodiment described above, and searches to see if a name written in a different way from the drug name (for example, the name of a drug listed in the drug price list) listed in the matched information is included in the common name link table 270 of Figure 36.

[0432] If the search reveals that the common name link table 270 contains names written in different ways, the MPU 210a will generate screen data for a literature search result screen including a literature information list in which the names written in different ways are arranged in the word field in addition to the drug names (for example, drug names listed in the drug price list) listed in the matched information, and send this screen data to the terminal device 30. The terminal device 30 will then receive this screen data and display it on the literature search result screen at S26 in the second flowchart or S42 in the third flowchart.

[0433] The literature search result screen displayed in this manner, compared to the literature search result screen 47 of the first embodiment shown in Figure 14, displays multiple different spellings for the same drug, which may include names written in spellings that the user is familiar with, making it easier to understand the contents of the drug and improving the usefulness of the final search result screen.

[0434] The processing of the third embodiment relating to the literature search result screen described above can also be applied to the second embodiment, as in the case of the first embodiment described above. If information containing both the drug name and side effect term of the search word is found at S139 of the fifth flow chart shown in Figure 32 of the second embodiment or S159 of the sixth flow chart in Figure 33, the MPU 110a performs the processing of the third embodiment described above and executes a process to search whether a name written differently from the drug name (e.g., the name of a drug listed in the drug price list) described in the found information is included in the generic name link table 270 of Figure 36.

[0435] If the generic name link table 270 contains a name with different spellings, screen data for a literature search result screen including a literature information list listing the name with different spellings in addition to the drug name (for example, the name of a drug listed in the drug price list) listed in the hit information is generated and transmitted to the terminal device 30. Based on this screen data, the terminal device 30 displays the literature search result screen at step S126 of the fifth flowchart or step S142 of the sixth flowchart. As a result, in the second embodiment as well, multiple different spellings for the same drug are listed on the drug name list screen, making it easier for the user to understand the contents of the drug.

[0436] Note that the example of applying the processing according to the third embodiment to the above-mentioned final search result screen has been explained in the case where the data group to be searched is big data 5 containing information on medical-related literature, but of course the processing according to the third embodiment can also be applied when the data group to be searched is other types of data group.

[0437] For example, the processing according to the third embodiment can also be applied when the data group to be searched is the X Hospital electronic medical record DB6 (see FIG. 2(b)), which is a data group storing patient information and managed by the X Hospital's electronic medical record system. In this case, the name of a drug (e.g., a drug name listed in the NHI drug price list) placed in the first column 22a on the electronic medical record search result screen (see electronic medical record search result screen 48 in FIG. 15) displayed will be displayed alongside the name of a drug in a different notation identified by the processing according to the third embodiment. This provides an advantage that, when identifying the drugs administered to patients (extracted patients) included in the electronic medical record search result screen, the drugs are displayed in various notations, allowing the user to confirm the drug name in a notation familiar to the user.

[0438] Similarly, the processing according to the third embodiment can also be applied when the data group to be searched is the Y Pharmacy's electronic medical history DB7 (see FIG. 2(c)), which is a data group storing a different type of patient information and is managed by the Y Pharmacy's electronic medical history system. In this case, the name of a drug (e.g., the name of a drug listed in the drug price list) placed in the first column 23a on the electronic medical history search result screen (see electronic medical history search result screen 49 in FIG. 16) is displayed, along with a different spelling of the name identified by the processing according to the third embodiment. As a result, when checking the drugs that have been administered to patients (extracted patients) included in the electronic medical history search result screen, the drugs are displayed using each spelling of the common name, making it possible to understand the drug by its name in a spelling that is familiar to the user.

[0439] The processing according to the third embodiment described above can also be applied to various modified examples of the first embodiment and various modified examples of the second embodiment, and by presenting the user with multiple different names of generic names for the same drug, it can contribute to making it easier for the user to recognize the drug. Note that the content of the first or second embodiment applies to all aspects other than those described above for the third embodiment.

[0440] Furthermore, the content of the above-mentioned third embodiment has been mainly explained in the case where, based on the name of a drug listed in the drug price list, other notified names, trade names (trade names), or names of different notations based on the drug name in the prescription computer processing system are identified from the common name link table 270 in Figure 36, but the content of the third embodiment can also be applied to identifying the remaining notified names based on the notified name, identifying the remaining notified names based on the trade names (trade names), or identifying the remaining notified names based on the notified name based on the drug name in the prescription computer processing system. [Example]

[0441] 37 is a schematic diagram showing the overall system configuration including an example of a side effect information retrieval system 300 according to a fourth embodiment (Example 4) of the present invention. The side effect information retrieval system 300 according to the fourth embodiment basically has some commonalities with the configuration according to the third embodiment described above, and is characterized in that it executes processing that can also handle practical handling of drug names.

[0442] Specifically, as explained in the third embodiment, the generic names, which are names commonly used as drug names by medical professionals, include the drug name listed in the drug price list, the notified name, the brand name (trade name), and the name of the drug in the prescription computer processing system, but in addition to these generic names, medical institutions (hospitals, clinics, pharmacies, etc.) may use their own names for drugs. Such drug names uniquely assigned by medical institutions (hereinafter referred to as unique names) often consist of a distinctive name portion of the generic name of the drug, or the name of an ingredient representing the drug's ingredient, combined with characters, etc., representing the drug's form, etc.

[0443] In the fourth embodiment, the above-mentioned unique names can also be handled, and a process is performed to identify the generic name (at least one of the drug name listed in the drug price list, the notified name, the brand name, or the drug name in the prescription computer processing system) from the unique name, and a search process, etc. is performed from the results of this process. In the fourth embodiment, the process of identifying the generic name from the unique name uses the technology disclosed in Japanese Patent Application No. 2019-129324 (hereinafter referred to as the disclosed technology).

[0444] Therefore, the server device 310 (side effect information search device) according to the fourth embodiment uses a unique na...

Claims

1. In a side effect information search system, a search process is performed using search words to search a data set containing information on drug names and side effect terms that indicate side effects of drugs corresponding to the drug names. a synonymous side effect table showing synonymous side effect terms for a side effect term, and including word combination side effect terms formed by combining multiple words among the synonymous side effect terms; a side effect word table indicating individual words used in combination with the word combination side effect terms included in the synonymous side effect table and the order in which each word is combined; a synonym table indicating synonymous relationships for the words indicated in the side effect word table; a side effect term specifying means for specifying synonymous side effect terms for the side effect terms to be searched based on the synonymous side effect table; a word specifying means for specifying, when the side effect term specifying means specifies a plurality of word combination side effect terms, individual words used in the combination for each of the specified plurality of word combination side effect terms based on the side effect word table; a synonymous word group specifying means for specifying a synonymous word group having a synonym relationship based on the synonymous word table for each of the individual words specified by the word specifying means; a generating means for extracting one word for each synonymous word group identified by the synonymous word group identifying means and combining the extracted words in the order of combination shown in the side effect word table to generate a side effect term; a means for performing a search process on the data group using the side effect terms generated by the generating means as search words; Equipped with the side effect terms generated by the generating means include side effect terms formed by combining a plurality of words identified by the side effect term identifying means and side effect terms with different spellings to the side effect terms used for searching, A side effect information search system, characterized in that the search words used in the search process include the side effect terms with different spellings.

2. In a side effect information search system, a search process is performed using search words to search a data set containing information on drug names and side effect terms that indicate side effects of drugs corresponding to the drug names. a synonymous side effect table showing synonymous side effect terms for a side effect term, and including word combination side effect terms formed by combining multiple words among the synonymous side effect terms; a side effect word table indicating individual words used in combination with the word combination side effect terms included in the synonymous side effect table and the order in which each word is combined; a synonym table indicating synonymous relationships for the words indicated in the side effect word table; a side effect term specifying means for specifying a word combination side effect term that becomes a synonymous side effect term based on the synonymous side effect table for a side effect term for search formed by combining a plurality of words; a word specifying means for specifying individual words to be used in combination for each of the word combination side effect terms specified by the side effect term specifying means and the side effect terms for search based on the side effect word table; a synonymous word group specifying means for specifying a synonymous word group having a synonym relationship based on the synonymous word table for each of the individual words specified by the word specifying means; a generating means for extracting one word for each synonymous word group identified by the synonymous word group identifying means and combining the extracted words in the order of combination shown in the side effect word table to generate a side effect term; a means for performing a search process on the data group using the side effect terms generated by the generating means as search words; Equipped with the side effect terms generated by the generating means include the word combination side effect terms identified by the side effect term identifying means and side effect terms with different spellings to the side effect terms used for search, A side effect information search system, characterized in that the search words used in the search process include the side effect terms with different spellings.

3. The side effect information search system of claim 1 or claim 2, wherein the generation means generates side effect terms by extracting one word for each synonymous word group identified by the synonymous word group identification means and combining them in the order of combination shown in the side effect word table, and generates side effect terms in all possible combinations.

4. In the side effect word table, for each word used in a word combination side effect term, a value level is indicated according to the importance in expressing the meaning of the word combination side effect term, the generating means is configured to generate a plurality of side effect terms; A side effect information search system as described in claim 1 or claim 2, wherein the means for performing search processing of the data group is configured to perform search processing by using each of the multiple side effect terms generated by the generation means as search words in an order according to the level values ​​of the words contained in each side effect term.

5. The data group includes literature information or patient information describing patient names as information describing drug names and side effect terms indicating side effects of the drug corresponding to the drug names, A drug link table showing drug names corresponding to side effect terms; a drug name identification means for identifying a drug name corresponding to the side effect term found in the search process based on the drug link table; a means for performing an output process of the drug name identified by the drug name identification means; a drug name receiving means for receiving any one of the drug names outputted; means for extracting, from the data group, literature information including information describing the side effect terms found in the search process and the drug names received by the drug name receiving means, or patient information describing the patient names; The side effect information search system according to claim 1 or 2, comprising:

6. It has a drug link table that shows drug names corresponding to side effect terms, 3. The side effect information search system according to claim 1, wherein the side effect terms included in the drug link table are used as the side effect terms for the search.

7. A drug link table showing drug names corresponding to side effect terms; a search term receiving means for receiving a side effect term for a search and a drug name for a search; A means for identifying a drug name corresponding to the side effect term found in the search process based on the drug link table; a drug name specifying means for specifying, among the specified drug names, a drug name that is the same as the drug name for search accepted by the search term accepting means; The side effect information search system according to claim 1 or 2, comprising:

8. The information contained in the data group further includes symptom terms indicating symptoms, a symptom table indicating individual words used in combination for symptom terms formed by combining multiple words; a symptom synonym table showing synonymous words that are synonymous with symptom words that are words that best represent the meaning of symptoms related to symptom terms among the words shown in the symptom table; a symptom word pronunciation table indicating pronunciation words representing pronunciations of the words indicated in the symptom table and the synonymous words indicated in the symptom synonym word table; a side effect symptom link table indicating side effect terms corresponding to the symptom terms included in the symptom table; a combined word specifying means for specifying, based on the symptom table, individual words used in a combination of a symptom term formed by combining a plurality of search words; a means for identifying synonymous words from the symptom synonym table for symptom words among the words identified by the combined word identifying means; means for identifying pronunciation words for each of the symptom words and the identified synonymous words based on the symptom word pronunciation table; a word search means for performing a search process to determine whether the symptom words, the synonymous words, and the pronunciation words are included in a data set; a symptom identification means for identifying symptom terms using words found to be included in the search process of the word search means based on the symptom table; a side effect specifying means for specifying a side effect term corresponding to the symptom term specified by the symptom specifying means based on the side effect symptom link table; Equipped with 3. The side effect information search system according to claim 1, wherein the side effect terms identified by the side effect identification means are used as the side effect terms for search.

9. The side effect symptom link table indicates a case where two or more symptom terms correspond to one side effect term, The side effect information search system of claim 8, further comprising a means for calculating, when there are multiple symptom terms corresponding to the side effect terms identified by the side effect identification means, the ratio between the number of symptom terms identified by the symptom identification means and the number of symptom terms corresponding to the side effect terms identified by the side effect identification means.

10. The side effect information search system according to claim 8, wherein symptom terms included in the symptom table are used as symptom terms for the search.

11. A drug link table showing drug names corresponding to side effect terms; a search term receiving means for receiving symptom terms and drug names for search; A means for identifying a drug name corresponding to the side effect term identified by the side effect identification means based on the drug link table; a drug name specifying means for specifying, from among the specified drug names, a drug name that is the same as the drug name for search accepted by the search term accepting means; a means for performing an output process of the drug name identified by the drug name identification means; The side effect information search system according to claim 8, comprising:

12. A side effect information search device that uses search words to search a data group containing information on drug names and side effect terms that indicate side effects of drugs corresponding to the drug names, a side effect term specification means for specifying a synonymous side effect term for a search side effect term based on a synonymous side effect table which indicates synonymous side effect terms for the side effect term and in which the synonymous side effect terms include word combination side effect terms formed by combining multiple words; When the side effect term identification means identifies a plurality of word combination side effect terms, a word identification means identifies each word used in combination for each of the plurality of word combination side effect terms based on a side effect word table that indicates each word used in combination for the word combination side effect terms included in the synonymous side effect table and the order of combination of each word; a synonymous word group specifying means for specifying a synonymous word group for each of the individual words specified by the word specifying means, based on a synonymous word table indicating synonymous relationships for the words specified in the side effect word table; a generating means for extracting one word for each synonymous word group identified by the synonymous word group identifying means and combining the extracted words in the order of combination shown in the side effect word table to generate a side effect term; a means for performing a search process on the data group using the side effect terms generated by the generating means as search words; Equipped with the side effect terms generated by the generating means include side effect terms formed by combining a plurality of words identified by the side effect term identifying means and side effect terms with different spellings to the side effect terms used for searching, The side effect information search device is characterized in that the search words used in the search process include the side effect terms with different spellings.

13. A side effect information search device that uses search words to search a data group containing information on drug names and side effect terms that indicate side effects of drugs corresponding to the drug names, a side effect term specification means for specifying a word combination side effect term that is a synonymous side effect term for a search side effect term formed by combining multiple words, based on a synonymous side effect table that indicates side effect terms that are synonymous with the side effect term, and the synonymous side effect terms include word combination side effect terms formed by combining multiple words; a word specifying means for specifying individual words to be used in combination for each of the word combination side effect terms specified by the side effect term specifying means and the side effect terms for search, based on a side effect word table indicating individual words to be used in combination for the word combination side effect terms included in the synonymous side effect table and the order of combination of each word; a synonymous word group specifying means for specifying a synonymous word group for each of the individual words specified by the word specifying means, based on a synonymous word table indicating synonymous relationships for the words specified in the side effect word table; a generating means for extracting one word for each synonymous word group identified by the synonymous word group identifying means and combining the extracted words in the order of combination shown in the side effect word table to generate a side effect term; a means for performing a search process on the data group using the side effect terms generated by the generating means as search words; Equipped with the side effect terms generated by the generating means include the word combination side effect terms identified by the side effect term identifying means and side effect terms with different spellings to the side effect terms used for search, The side effect information search device is characterized in that the search words used in the search process include the side effect terms with different spellings.

14. The information contained in the data group further includes symptom terms indicating symptoms, a combined word specifying means for specifying individual words used in combination for a symptom term formed by combining a plurality of search words, based on a symptom table indicating individual words used in combination for a symptom term formed by combining a plurality of words; a means for identifying synonymous words for a symptom word among the words identified by the combined word identifying means from a symptom synonym table that indicates synonymous words for a symptom word that best represents the meaning of a symptom related to a symptom term among the words indicated in the symptom table; a means for identifying pronunciation words for each of the symptom words and the identified synonymous words based on a symptom word pronunciation table that indicates pronunciation words for the words indicated in the symptom table and the synonymous words indicated in the symptom synonym table; a word search means for performing a search process to determine whether the symptom words, the synonymous words, and the pronunciation words are included in a data set; a symptom identification means for identifying symptom terms using words found to be included in the search process of the word search means based on the symptom table; a side effect specifying means for specifying a side effect term corresponding to the symptom term specified by the symptom specifying means based on a side effect symptom link table indicating a side effect term corresponding to the symptom term included in the symptom table; Equipped with The side effect information search device according to claim 12 or 13, wherein the side effect terms identified by the side effect identification means are used as the side effect terms for search.

15. A side effect information search method in which a side effect information search device searches a data group including information describing drug names and side effect terms indicating side effects of drugs corresponding to the drug names, using search words: The side effect information search device includes: a step of identifying a synonymous side effect term for the search side effect term based on a synonymous side effect table that indicates synonymous side effect terms for the side effect term and includes word combination side effect terms formed by combining multiple words, a word specifying step of specifying, when a plurality of word combination side effect terms are specified in the specifying step, individual words used in combination for each of the plurality of word combination side effect terms based on a side effect word table indicating individual words used in combination for the word combination side effect terms included in the synonymous side effect table and the order of combination of each word; a synonymous word group identification step of identifying a synonymous word group having a synonymous relationship with each of the individual words identified in the word identification step, based on a synonymous word table indicating synonymous relationships with the words indicated in the side effect word table; a generating step of extracting one word for each synonymous word group identified in the synonymous word group identifying step, and combining the extracted words in the order of combination shown in the side effect word table to generate a side effect term; performing a search process on the data group using the side effect terms generated in the generating step as search words; Run The side effect terms generated in the generating step include side effect terms of multiple word combinations identified in the identifying step and side effect terms with different spellings to the side effect terms used for searching, A side effect information search method, characterized in that the search words used in the search process include the side effect terms with different spellings.

16. A side effect information search method in which a side effect information search device searches a data group including information describing drug names and side effect terms indicating side effects of drugs corresponding to the drug names, using search words: The side effect information search device includes: a step of identifying a word combination side effect term that is a synonymous side effect term for a search side effect term formed by combining multiple words, based on a synonymous side effect table that indicates a synonymous side effect term for the side effect term and includes a word combination side effect term formed by combining multiple words among the synonymous side effect terms; a word specifying step of specifying individual words to be used in combination for each of the word combination side effect terms specified in the specifying step and the side effect terms for search, based on a side effect word table indicating individual words to be used in combination for the word combination side effect terms included in the synonymous side effect table and the order of combination of each word; a synonymous word group identification step of identifying a synonymous word group having a synonymous relationship with each of the individual words identified in the word identification step, based on a synonymous word table indicating synonymous relationships with the words indicated in the side effect word table; a generating step of extracting one word for each synonymous word group identified in the synonymous word group identifying step, and combining the extracted words in the order of combination shown in the side effect word table to generate a side effect term; performing a search process on the data group using the side effect terms generated in the generating step as search words; Run The side effect terms generated in the generating step include side effect terms with different spellings than the word combination side effect terms identified in the identifying step and the side effect terms used for searching, A side effect information search method, characterized in that the search words used in the search process include the side effect terms with different spellings.

17. A computer program for causing a computer to execute a search process for a data group including information describing drug names and side effect terms indicating side effects of the drug corresponding to the drug name, using a search word, The computer, a step of identifying a synonymous side effect term for the search side effect term based on a synonymous side effect table that indicates synonymous side effect terms for the side effect term and includes word combination side effect terms formed by combining multiple words, a word specifying step of specifying, when a plurality of word combination side effect terms are specified in the specifying step, individual words used in combination for each of the plurality of word combination side effect terms based on a side effect word table indicating individual words used in combination for the word combination side effect terms included in the synonymous side effect table and the order of combination of each word; a synonymous word group identification step of identifying a synonymous word group having a synonymous relationship with each of the individual words identified in the word identification step, based on a synonymous word table indicating synonymous relationships with the words indicated in the side effect word table; a generating step of extracting one word for each synonymous word group identified in the synonymous word group identifying step, and combining the extracted words in the order of combination shown in the side effect word table to generate a side effect term; performing a search process on the data group using the side effect terms generated in the generating step as search words; It is designed to execute The side effect terms generated in the generating step include side effect terms of multiple word combinations identified in the identifying step and side effect terms with different spellings to the side effect terms used for searching, A computer program characterized in that the search words related to the search process include the side effect terms with different spellings.

18. A computer program for causing a computer to execute a search process for a data group including information describing drug names and side effect terms indicating side effects of the drug corresponding to the drug name, using a search word, The computer, a step of identifying a word combination side effect term that is a synonymous side effect term for a search side effect term formed by combining multiple words, based on a synonymous side effect table that indicates a synonymous side effect term for the side effect term and includes a word combination side effect term formed by combining multiple words among the synonymous side effect terms; a word specifying step of specifying individual words to be used in combination for each of the word combination side effect terms specified in the specifying step and the side effect terms for search, based on a side effect word table indicating individual words to be used in combination for the word combination side effect terms included in the synonymous side effect table and the order of combination of each word; a synonymous word group identification step of identifying a synonymous word group having a synonymous relationship with each of the individual words identified in the word identification step, based on a synonymous word table indicating synonymous relationships with the words indicated in the side effect word table; a generating step of extracting one word for each synonymous word group identified in the synonymous word group identifying step, and combining the extracted words in the order of combination shown in the side effect word table to generate a side effect term; performing a search process on the data group using the side effect terms generated in the generating step as search words; It is designed to execute The side effect terms generated in the generating step include side effect terms with different spellings than the word combination side effect terms identified in the identifying step and the side effect terms used for searching, A computer program characterized in that the search words related to the search process include the side effect terms with different spellings.

Images