JPH0999039A - Chemical interaction check method - Google Patents

Abstract

(57) [Abstract] [PROBLEMS] To automatically generate a drug interaction table indicating a combination of drugs that cause an interaction, and to rapidly combine a drug that causes an interaction by using the drug interaction table. To provide a drug interaction check method capable of being detected. In a drug interaction checking method for checking the presence or absence of a drug combination that causes an interaction with respect to a plurality of drugs, a combination of all drugs is generated for a plurality of inputted drugs, and a combination of the drugs is generated. Select a word string consisting of high-order word, low-order word, synonym, and different notation word of one drug name from the dictionary, and use the other drug name in the combination as a key in the character string data searched from the drug information database for the above word A drug interaction checking method characterized in that it is configured to check whether or not a column is included by a character search.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a drug interaction checking method suitable for supporting a drug prescription issuance and a drug prescription inspection performed by a doctor or pharmacist at a hospital or pharmacy.

[0002]

2. Description of the Related Art When prescribing a drug, a doctor or pharmacist considers whether the dose of the drug is appropriate, or whether there is a combination of drugs that cause an interaction with the prescribed drug. This information is described in the efficacy statement attached to the drug, and the doctor or pharmacist will prescribe the drug based on the efficacy statement. However, the number of drugs on the market is enormous, and it takes a lot of time and effort to check the dose and interaction of all the prescribed drugs based on the written efficacy statement. Doctors and pharmacists know the content of the efficacy report for a certain number of drugs, and do not check the content of the efficacy report for prescribed drugs one by one, but drugs and new drugs that are not usually handled often I have no choice but to refer to the effect book. Especially regarding the interaction
Combinations of drugs that are prohibited to be used in combination, or one drug may increase the side effects of the other drug or the efficacy may be diminished by the combined use. There are countless numbers, and it is almost impossible to grasp all of them. For this reason, a drug interaction table is created in advance by listing the drug names that cause interactions with each drug from the efficacy statement for each drug.
To see if there are any combinations of chemicals that
An interaction check method has been proposed which supports issuing of prescriptions and auditing of prescriptions by checking all the combinations of prescribed drugs with a computer using this table. Regarding this, it is possible to refer to the description of "Development of drug interaction information retrieval system" by Kusunoki et al. (Abstracts of the 14th Joint Meeting of Medical Informatics, pages 227-228 (Nov., 1994)).

[0003]

However, it takes a lot of labor and time to manually create the drug interaction table for a huge number of drugs. In addition, new drugs will appear one after another, so the table above must be maintained continuously. In addition, in the efficacy statement, the drug that causes the interaction is not the drug name, but the generic name assigned to each drug group of the same main component or the drug efficacy name assigned to the drug group having the same pharmacological action. It is described in. Usually, the prescription is given by the drug name or the product name given by the pharmaceutical company that provides the drug. Therefore, the product name for determining which generic name or pharmacological drug efficacy group the prescribed drug belongs to. A thesaurus consisting of drug name, generic name, and pharmacological effect name is also required. Therefore, in order for the interaction check system to operate properly, it is also necessary to always check the consistency between the registered drug name in the interaction table and the thesaurus. The present invention has been made in view of the above circumstances, and an object thereof is to solve the above-mentioned problems in the conventional art and automatically generate a drug interaction table indicating a combination of drugs that cause an interaction. Another object of the present invention is to provide a drug interaction checking method that enables the drug interaction table to be detected at high speed by using the drug interaction table.

[0004]

The above object of the present invention is to provide a drug interaction checking method for checking the presence or absence of a combination of drugs that cause an interaction with respect to a plurality of drugs. A combination of all medicines is generated, a word string consisting of a higher word, a lower word, a synonym, and a different notation word of one medicine name in the combination is selected from the dictionary, and the medicine name of the other medicine in the combination is used as a key. This can be achieved by a drug interaction checking method characterized by checking whether or not the word string is included in the character string data searched from the information database by a character search. The drug interaction checking method according to the present invention comprises a drug interaction checking unit 101 and a drug information maintenance unit 102. In the drug information maintenance unit, an attached document text file 106 in which an attached document of each drug is converted into text code data is input. The Hiragana notation is deleted for each drug, and the condensed text in which the words included in the predetermined dictionary are deleted is linked to the drug name and the attached text file of the drug, and the drug interaction table 105 To generate. The drug interaction check unit generates all "two drug pairs" for a plurality of input drugs, and in each pair, the drug classification thesaurus 103 and synonyms for one drug name. Different notation dictionary 1
The synonyms, different expressions, upper terms, and lower terms are obtained from 04, and when the synonyms, different expressions, upper terms, and lower terms have no hiragana, the above condensation in the record of the drug interaction table for the other drug name. It is checked by performing a character search whether or not the above-obtained synonyms, different expressions, upper terms, and lower terms are included in the text, and the combination of the drugs that cause the interaction is verified. If the upper word and the lower word have hiragana, the above-mentioned synonym, different notation, upper word, and lower word are included in the above attached document text file in the record of the drug interaction table for the other drug name. It is characterized by checking whether or not it is present by character search and verifying the combination of drugs that cause an interaction.

[0005]

BEST MODE FOR CARRYING OUT THE INVENTION In the drug interaction checking method according to the present invention, a drug interaction table is automatically generated from an attached document text file containing the efficacy documents of each drug using a predetermined procedure and a dictionary. Since it is created from the condensed text created and the attached text file itself,
It is possible to generate a drug interaction table with minimal human intervention. Here, the condensed text includes a drug name that causes an interaction with the drug. Next, in the drug interaction check unit, synonyms, different expressions, broader terms of one drug, and two words of the drug pair,
Subordinate terms are the drug classification thesaurus and synonyms / different notation dictionary,
Obtained from the unnecessary word dictionary, this synonym, different notation, broader terms,
A character search is used to check if the subordinate word is included in the condensed text or package insert text file of the other drug. At this time, if the synonyms, different expressions, high-rank words, and low-rank words do not include Hiragana notation, character search is performed using condensed text instead of the original attached text file, so the search time is shortened. It is possible to detect a combination of drugs that cause an action at high speed. If the synonyms, different notations, upper terms, and lower terms include hiragana notation, it is unavoidable that character search processing will be performed on the original attached document text file. But,
Many drug names are usually written in kanji and katakana, and it can be considered that they rarely occur when character search processing is performed on the original attached text file, so drug combinations that cause interactions can be detected at high speed. can do. Hereinafter, embodiments of the present invention will be described in more detail with reference to the drawings.

[0006]

EXAMPLE A system configuration according to an embodiment of the present invention will be described with reference to FIG. The system according to this embodiment includes a drug interaction check unit 101 and a drug information maintenance unit 102. Chemical interaction check unit 101
When a doctor or pharmacist prescribes a drug for a patient, whether there is a problem in the combination of the drug prescribed in the past days and the drug prescribed this time,
It is a subsystem for checking using the medicine classification thesaurus 103, synonym / different notation dictionary 104, medicine interaction table 105, and unnecessary word dictionary (not shown) described below.

The drug information maintenance unit 102 is a subsystem that automatically creates the drug interaction table 105. In the drug information maintenance unit 102, an attached document text file 106 in which an attached document in which efficacy, side effects, interactions, etc. of each drug are described is converted into text code data is input, and a drug classification thesaurus 103, synonyms, The drug interaction table 105 is automatically generated using the different notation dictionary 104 and the unnecessary word dictionary. The drug classification thesaurus 103 is a development of a thesaurus for each drug, which is a hierarchical classification of drugs such as trade names, drug names, generic names, and pharmacological drug efficacy names. The synonym / different notation dictionary 104 is a collection of synonyms and different notations for the brand name, drug name, generic name, and pharmacological effect name of each drug. Drug interaction table 10
5 shows a summary of drugs that should be carefully administered in combination and drugs that should be avoided in combination. Further, the unnecessary word dictionary includes, for example, “administration”, “observation”,
It includes a word such as “enhancement” that is not a drug name even though it is a medical-related word.

Drugs that should be carefully administered for concomitant use and drugs that should be avoided for concomitant use should be registered under the names of which level of brand name, drug name, generic name, or pharmacological effect name. It doesn't matter. The detailed structure of the drug interaction table 105 is as shown in FIG. 2, and each record includes a drug name 201, concomitant drug condensation information 202, and concomitant drug original information 203. Drug interaction table 10
In the item of 5, the drug name 201 and the concomitant drug original information 203 only need to be embedded in the part corresponding to each drug from the attached document text file 106. Therefore, the method of generating the concomitant drug condensing information 202 will be described below.
This will be described in detail with reference to FIG. The attached document text file 106 is provided, for example, by the Japan Medical Information System Development Center (MEDIC-DC).

Step 301: For each drug, extract the part relating to the interaction of the combined drug in the attached document text file 106. Step 302: With respect to the portion related to the concomitant drug extracted in step 301, delete the hiragana notation portion in the character string to create a condensed text that is decomposed into a word string. Step 303: For each word of the condensed text created in step 302, check whether or not it is a word registered in the unnecessary word dictionary, delete it if it is a word registered in the unnecessary word dictionary, and if not, convert it to a condensed text Create condensed text that has been retained.

Step 304: With respect to the words left undeleted in step 303, it is checked whether or not the words are registered in any of the medicine classification thesaurus and the synonym / different notation dictionary, and unnecessary words are registered in the unregistered words. Spit to the candidate file. Step 305: A user who maintains the drug information checks the words in the unnecessary word candidate file, and registers the drug name in the drug classification thesaurus and synonym / different notation dictionary if the drug name. If it is not a drug name, register it in the unnecessary word dictionary, and step 3
Delete the word from the condensed text obtained in 03.
If the drug classification thesaurus and the synonym / different notation dictionary are complete, the drug interaction table 105 is automatically generated by the processing of steps 301 to 305 described above.

Next, the chemical interaction maintenance unit 1 described above.
The operation of the drug interaction check unit using the drug interaction table 105 generated in 02 will be described in detail with reference to FIG. Here, it is assumed that the drug name at the time of prescription is the drug name. Step 401: The medicine name prescribed to the patient with ID = n is set in the array medn (i). Step 402: The name of a drug prescribed in the past for a predetermined number of days is detected from the prescription file of patient ID = n and added to the array medn (i). Step 403: Generate combinations c (k) = (medn (i), medn (j)) (i ≠ j) of all drug names from the drug names of the array medn (i). Step 404: For each drug name medn (i), the product name, drug name, generic name, pharmacological drug effect name from the drug classification thesaurus 103, and the product name from the synonym / different notation dictionary 104,
Synonyms and different expressions for each drug name, generic name, and pharmacological drug effect name are extracted and set in the array medv (i, j).

Step 405: Each combination of drug names generated in the above step 403 c (k) = (medn (i), medn (j))
On the other hand, if medn (i) does not include the Hiragana notation, the drug name medn (i) in the drug interaction table 105.
From the record, it is checked for each l whether medv (j, l) created in step 404 is included in the item of concomitant drug condensation information 202. If medn (i) contains the Hiragana notation, the drug interaction table 105
Of the medicine name medn (i) of medv (j, j
Check for each l whether or not l) is included. At this time, if included, the character string included in the item of the combination drug original information 203 is extracted from the record of the drug name medn (i) in the drug interaction table,
It is displayed on the display of the operating terminal and issues a warning. Step 406: The same process as step 405 is performed by medn
(i) and medn (j) are exchanged.

According to the above-described embodiment, it is possible to automatically generate a drug interaction table indicating a combination of drugs that cause an interaction, and use the drug interaction table to speed up the combination of drugs that cause an interaction. There is an effect that it is possible to realize a drug interaction check method that can be detected. Next, as another embodiment, operations of the drug interaction check unit 101 and the drug information maintenance unit 102 when the structures of the drug classification thesaurus 105 and the synonym / different notation dictionary 104 are changed will be described. In the present embodiment, the medicine classification thesaurus 602
As shown in FIG. 6, the compound name, the drug name, the generic name, and the pharmacological effect name are decomposed at the transition of kanji, katakana, hiragana, and alphanumeric characters, and the word in kanji is used as a drug. , A drug, a system, hydrochloric acid, inhibition, anti-Kanji, which is often used in drug names, and a dictionary 601 in which Kanji that expresses a pharmacological action are registered, are prepared in the form of word decomposition. In a similar way,
Synonym / different notation dictionary 6 that decomposes synonyms and different notations into words
03 is prepared. When the Kanji notation is decomposed using the dictionary 601, the longest matching method is adopted, which matches the word registered in the dictionary and has the longest word length. An automatic generation method of the drug interaction table at this time will be described with reference to FIG.

Step 501: For each drug, the part of the package insert text file 106 relating to the interaction of the combined drug is extracted. Step 502: With respect to the portion related to the concomitant drug extracted in step 501, the portion written in hiragana in the character string is deleted to create condensed text that is decomposed into a word string. Step 503: For each word of the condensed text of step 502, it is checked whether or not it is a word registered in the unnecessary word dictionary, and if it is a registered word, it is deleted. create. Step 504: With respect to the remaining words that are not deleted in step 503, the words are decomposed at the transition of the notation and divided into word strings, and further, the dictionary 601 decomposes the kanji notation by the longest matching method to generate a text file.

Step 505: It is checked whether or not the word in the text file generated in step 505 is a word registered in the medicine classification thesaurus 602 and the synonym / different expression dictionary 603. The word before being decomposed in step 504 including the character string as a partial character string is discharged to the unnecessary word candidate file. Step 506: The user who maintains the drug information checks the words in the unnecessary word candidate file, and if it is a drug name, the drug classification thesaurus 602, the synonym / different notation dictionary 603.
Register with. If it is not a medicine name, it is registered in the unnecessary word dictionary and the word is deleted from the text file obtained in step 504.

In the medicine classification thesaurus 602 and the synonym / different notation dictionary 603 having the above-mentioned configuration, for example,
Synonyms such as "central nervous system drug" and "central nervous system depressant" that partially include the same character string are no longer regarded as different words. Therefore, even when the medicine classification thesaurus and the synonym / different expression dictionary are not very complete, the number of words discharged to the unnecessary word candidate file is reduced, and the manual processing is reduced. However, it has no effect on expressions such as "a drug containing a large amount of magnesium and calcium."

Next, the drug interaction checking unit 101 using the drug interaction table created in steps 501 to 505 will be described with reference to FIG. Step 701: The drug name prescribed for the patient with ID = n is set in the array medn (i). Step 702: A drug name previously prescribed for a predetermined number of days is detected from the prescription file of patient ID = n and added to the array medn (i). Step 703: Generate a combination c (k) = (medn (i), medn (j)) (i ≠ j) of all drug names from the drug names of the array medn (i).

Step 704: For each drug name medn (i), the product name, drug name, generic name, pharmacological drug effect name and its decomposition form from the drug classification thesaurus 602, and the product name from the synonym / different notation dictionary 603. , Synonyms and different notations for each drug name, generic name, and pharmacological effect name and their decomposed forms are taken out and set in the array medv (i, j). Step 705: For each combination of drug names c (k) = (medn (i), medn (j)) generated in step 703 above, if medn (i) does not include the hiragana notation From the record of the medicine name medn (i) in the medicine interaction table 105, it is checked for each l whether or not the item of the combined medicine condensation information 202 includes medv (j, l) created in step 704. . If medn (i) includes the Hiragana notation, the drug name medn in the drug interaction table 105.
From the record (i), it is checked for each l whether or not medv (j, l) created in step 704 is included in the item of the concomitant drug original information 202. At this time,
If included, the drug name from the drug interaction table
Original information on concomitant drugs from the record of medn (i) 20
The character string included in item 3 is taken out and displayed on the display of the operation terminal to issue a warning.

Step 706: The same processing as step 705 is performed by exchanging medn (i) and medn (j). According to the above-described embodiment, it is possible to automatically generate a drug interaction table indicating a combination of drugs that cause an interaction, and use the drug interaction table to quickly detect a combination of drugs that cause an interaction. This has the effect of realizing a possible drug interaction checking method. In addition, the above embodiment is an example,
It goes without saying that the present invention should not be limited to this.

[0020]

As described above in detail, according to the present invention, the following effects can be obtained. (1) In the drug interaction check section, the drug interaction table is automatically created using the prescribed procedure and dictionary from the attached document text file containing the efficacy documents for each drug, and the attached document text. Since it is created from the file itself, the drug interaction table can be created with minimal human intervention. (2) In the drug interaction check unit, for each drug pair, the synonym, different notation, high-order word, and low-order word for one drug are obtained from the drug classification thesaurus and the synonym / different notation dictionary. Use a character search to check whether the different notation, broader terms, and lower terms are included in the condensed text or package insert text file of the other drug. At this time, if the synonyms, different expressions, high-rank words, and low-rank words do not include Hiragana notation, character search is performed using condensed text instead of the original attached text file, so the search time is shortened. It is possible to detect a combination of drugs that cause an action at high speed. Synonyms, different notations, broader terms,
When the lower word includes the hiragana notation, it is unavoidable to perform the character search process on the original attached text file. However, many drug names are usually written in kanji and katakana, and it can be considered that this rarely occurs when character search processing is performed on the original attached text file. Is likely to be detected.

[Brief description of drawings]

FIG. 1 is a system configuration diagram according to an embodiment of the present invention.

FIG. 2 is a diagram illustrating a drug interaction table.

FIG. 3 is a diagram illustrating a method of generating a drug interaction table.

FIG. 4 is a diagram illustrating a drug interaction checking method.

FIG. 5 is a diagram illustrating a method of generating a drug interaction table.

FIG. 6 is a diagram illustrating a drug classification thesaurus and a synonym / different notation dictionary.

FIG. 7 is a diagram illustrating a drug interaction checking method.

[Explanation of symbols]

 101 drug interaction check unit 102 drug information maintenance unit 103 drug classification thesaurus 104 synonym / different notation dictionary 105 drug interaction table 201 drug name 202 concomitant drug condensation information 203 concomitant drug original information 601 dictionary 602 drug classification thesaurus 603 synonym / Different notation dictionary

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Akihide Hashizume 12 890 Kashimada, Saiwai-ku, Kawasaki-shi, Kanagawa 12 Hitachi, Ltd. Information Systems Division

Claims (6)

[Claims] 1. A drug interaction check method for checking the presence or absence of a drug combination that causes an interaction with respect to a plurality of drug types, wherein all drug combinations are generated for a plurality of input drug types, and the combination is generated. In the dictionary, a word string consisting of the upper word, the lower word, the synonym, and the different notation of one drug name in is selected from the dictionary, and the other drug name in the combination is used as a key to the character string data retrieved from the drug information database. A drug interaction checking method characterized by checking whether or not a word string is included by a character search. 2. The character string data retrieved from the drug information database using the other drug name in the drug combination as a key is the text code data itself including the drug information or the text code including the drug information. The condensed text is created by deleting words included in the unnecessary word dictionary from the data.
The drug interaction check method described. 3. The drug interaction checking method according to claim 2, wherein the unnecessary word dictionary includes the entire hiragana. 4. In the case where a word string consisting of an upper word, a lower word, a synonym, and a different notation word of one drug name in the drug combination includes hiragana notation, the other drug name in the combination. The text code data itself including the information of the medicine is selected as the character string data retrieved from the medicine information database with the key as a key, and when the hiragana notation is not included, the character string data is used as the character string data. The condensed text described is selected, and it is checked whether or not a word string consisting of an upper word, a lower word, a synonym, and a different notation word of one drug name in the drug combination is included. The drug interaction check method described in 1. 5. A word string consisting of an upper word, a lower word, a synonym, and a different notation word of one drug name in the drug combination is selected from a dictionary, and the other drug name in the drug combination is used as a key. It is checked by a character search whether or not the word string is included in the character string data retrieved from the information database, and if it is included, the text code data including the drug information according to claim 2 is displayed on the display device. A method for checking drug interaction, which comprises: 6. When the condensed text is created by deleting the words contained in the unnecessary word dictionary from the text code data including the drug information according to claim 2, the condensed text word is the word according to claim 1. It is checked whether or not it is included in the dictionary used when selecting a word string consisting of an upper word, a lower word, a synonym, and a different notation word of one drug name in each drug combination, and if it is included in the dictionary If the word is not included and the drug name is registered, the word is registered in the dictionary, and if it is not included and the drug name is not included, the word is registered in the unnecessary word dictionary and the word is the condensed text. The method for checking drug interaction according to claim 1, further comprising a procedure of deleting the drug interaction from the drug.