JP5119821B2 - Synonymous disease name selection device - Google Patents

Description

  The present invention relates to a synonymous disease name selection device and a synonymous disease name selection program for selecting a disease name having a different notation expressing a synonymous disease.

  In the medical field, unification of disease name notation is an important issue. If disease names are standardized nationwide, it will be possible to conduct demographic surveys on disease occurrence by gender, age group, etc., and obtain useful statistical information. In addition, even in the clinical field, if the disease names are unified, it is possible to easily refer to past histories such as examinations, treatments, and drug administrations for diseases across hospitals. Can also reduce miscommunication between hospitals.

Therefore, various methods have been proposed in order to unify different notations expressing synonymous diseases. For example, a method has been proposed in which an input disease name is divided into main disease names and modifiers, and these are standard-coded using a standard code dictionary (see Patent Document 1). In addition, a method for specifying a disease name notation expressing the same disease name from the similarity of the character string of the disease name has been proposed (see Non-Patent Document 1).
JP 2004-220167 A "Examination of method to transfer disease name data to new disease name master", Proceedings of the 22nd Medical Informatics Conference, 2003

  In the medical field, new concepts and technical terms that express them are constantly being created as technology advances. With regard to diseases, with the advancement of medical research such as improvement of testing techniques and genome analysis, new disease concepts are always discovered, and new disease names are created. The notation of the disease name that expresses the concept of such a new disease is largely shaken, and quite different notation is often used between hospitals.

  Here, in the technique according to Patent Document 1, the disease name is divided and the identity of the disease name is determined by using the corresponding code. However, in the report or medical record described in the medical field, the corresponding code is used. Disease names and modifiers that do not exist are often used, and there are many disease name notations that cannot be encoded. In addition, the provision of a code for a disease is a work with a large time cost, and there is no code corresponding to at least a certain period for a new disease concept. Moreover, although the technique according to Non-Patent Document 1 uses the similarity of the disease name character string, medical reports, medical records, and the like indicate “hepatoma” as “HCC”. The disease name notation that expresses is not necessarily composed of similar character strings.

  The present invention has been made in view of the above-described conventional circumstances, and is a disease name notation that is not coded as in the novel disease concept, and even if the disease name notation is completely different from the surface character string. The purpose of this study is to propose a technique that can select disease names with different notations expressing synonymous diseases.

  The present invention according to claim 1 includes: a first storage unit that stores diagnostic information including a patient's disease name and a date and time related to the diagnosis; and a plurality of pieces of diagnostic information stored in the first storage unit. Based on the acquisition means for acquiring the appearance frequency of each disease name for each predetermined period, the tendency of the temporal change of the appearance frequency related to one disease name and the tendency of the temporal change of the appearance frequency related to another disease name A synonymous disease name, comprising: a determination means for determining similarity between the two and an output means for selecting and outputting as synonyms disease names determined to have similar trends in temporal changes in appearance frequency It is a selection device.

  According to a second aspect of the present invention, there is provided the synonymous disease name selecting device according to the first aspect, wherein the synonymous disease name selecting device stores synonymous disease name information in which synonymous disease names are associated with each other in advance. Storage means, and the determination means excludes the disease names associated by the synonymous disease name information from being determined.

  According to a third aspect of the present invention, in the synonymous disease name selection device according to the second aspect, the output means outputs synonymous disease names to the second storage means, and stores them as synonymous disease name information. It is characterized by making it.

  According to a fourth aspect of the present invention, in the synonymous disease name selection device according to any one of the first to third aspects, the diagnostic information includes identification information of a diagnostic subject relating to a diagnosis of a patient. The determination means excludes disease names related to a common diagnosis subject from being determined.

  The invention of claim 5 is the synonymous disease name selection device according to any one of claims 1 to 4, wherein the diagnosis information includes identification information of a hospital that diagnosed a patient, The synonymous disease name selection device further includes a third storage unit that stores distance information related to a distance between hospitals, and the determination unit sequentially determines a combination of hospitals having a short distance. .

  The present invention according to claim 6 is a first storage means for storing diagnostic information including a disease name of a patient and a date and time related to the diagnosis, and a plurality of computers stored in the first storage means. Based on the diagnosis information, acquisition means for acquiring the appearance frequency of each disease name for each predetermined period, the tendency of the temporal change of the appearance frequency related to one disease name, and the temporal change of the appearance frequency related to another disease name A synonymous disease name selection program for functioning as an output unit for selecting and outputting synonyms of disease names determined to have similar temporal trends in appearance frequency, and a determining unit that determines similarity to the tendency of It is.

  According to the synonymous disease name selection device according to claim 1, even if the disease name notation is not coded as in the novel disease concept and the disease name notation is completely different from the surface character string, Based on the similarity of the tendency of the temporal change in the appearance frequency related to the disease, it is possible to select an illusory disease name expressing a synonymous disease. In particular, in the case of a new epidemic disease that is difficult to unify, the temporal probability distribution of patients with the same disease visiting the hospital will be constant regardless of the hospital, as long as it is limited to hospitals of the same size in the same region. Therefore, it is effective for the selection of names of allopathic diseases for such new epidemic disease concepts.

  According to the synonymous disease name selection apparatus according to claim 2, the combination of known synonymous disease names is excluded from the determination target, and the processing burden related to the selection of synonymous disease names is reduced.

  According to the synonymous disease name selection apparatus according to claim 3, the combination of the disease names determined to be synonymous is excluded from the subsequent determination targets, and the processing burden relating to the subsequent selection of the synonymous disease names is reduced.

  According to the synonymous disease name selection apparatus according to claim 4, combinations of disease names related to a common diagnostic entity (such as a hospital department or section or a doctor) considered to use a single disease name are excluded from the determination. The processing burden related to the selection of synonymous disease names is reduced.

  According to the synonymous disease name selection device according to claim 5, since it is considered that the tendency of the temporal change of the appearance frequency of a certain case in a close region is very similar, by judging in order from the close hospital, The accuracy of selecting synonymous disease names can be increased.

  According to the synonymous disease name selection program according to the sixth aspect, the synonymous disease name selection device having the above-described effects can be realized using a computer.

The present invention will be specifically described based on an embodiment exemplified below.
In addition, when considering the departments in the hospital as a unit, it is considered that frequent communication between doctors such as reporting patient treatment / test progress at conference meetings etc., so in the following explanation, It is assumed that disease name notation is almost unified in a single hospital department.

FIG. 1 shows a functional block diagram of the synonymous disease name selection apparatus according to this example.
The synonymous disease name selection device of this example includes a document storage means 1, a synonymous disease name dictionary holding means 2, a disease name extraction means 3, a disease name appearance distribution calculation means 4, a synonymous disease name determination means 5, and a synonymous disease name output means 6. It is equipped with.

  The document storage unit 1 stores diagnostic information including a patient's disease name and a diagnosis date and time related to the diagnosis. In this example, medical reports such as medical reports and electronic medical records created by a plurality of hospital departments are stored, and the medical department ID, patient ID, disease name, document creation date and time (diagnosis date and time) ) Etc. are included. It is intended for medical documents of departments that share the same medical field (internal medicine, neurology, etc.) in different hospitals. For example, medical documents of internal medicine A in one hospital and internal medicine B in another hospital are stored. Has been.

  The synonymous disease name dictionary holding means 2 stores synonymous disease name information in which synonymous disease names are associated with each other in advance. In this example, as shown in the data example in FIG. 2, known synonymous disease names are held for each disease concept. For example, “stomach cancer” “stomach cancer” “stomach cancer” “stomach cancer” “stomach Ca.” It can be seen that represents a synonymous disease concept.

  The disease name extraction unit 3 extracts disease names from a plurality of medical documents stored in the document storage unit 1. In this example, a character string described in the field is extracted as a disease name from a structured medical document in which a field of “diagnosis name” exists. For example, an unstructured free text In the case of such a medical document, a disease name may be extracted using a proper noun extraction technique as disclosed in Japanese Patent Laid-Open No. 06-019959.

The disease name appearance distribution calculating means 4 acquires the appearance probability for each disease name for each predetermined period for each hospital department.
In this example, it is assumed that a medical document is created for each treatment day in a series of medical treatments for a certain patient, and a representative disease name (from a disease name and a diagnosis date / time described in each medical document) After the identification of the confirmed disease name) and the diagnosis date and time (first visit date and time), the appearance frequency is calculated based on the confirmed disease name and the first visit date and time.

  Specifically, medical documents related to the same patient in the same hospital department are arranged in chronological order (in order of creation date), and the medical document in which the disease name is confirmed in this medical document series (confirmed medical document) The disease name described in the confirmed medical document is extracted as the confirmed disease name, and the creation date and time of the first document in the medical document series is extracted as the first visit date and time. The medical document handled in this example has a field of “definite diagnosis name”, and the medical document in which the disease name is written in the field is determined as the definitive medical document, and the definite disease name is extracted. However, in the case of a free text medical document, for example, a pattern character string for extracting a definite disease name such as “It is a finding of XXX” is prepared, and a disease name matching this is extracted as a definite disease name. . If the definite disease name cannot be specified as described above, the last disease name appearing in the medical document series may be set as the definite disease name.

When the confirmed disease name and the first visit date and time are specified as described above, the disease name appearance distribution calculating means 4 generates the appearance of each confirmed disease name for each predetermined period (such as weekly or monthly) for each department of each hospital. The probability (ratio of the appearance frequency of each confirmed disease name to the total appearance frequency of all confirmed disease names that appear in the same period) is calculated.
When the onset date and time of the symptom is included in the medical document, the onset date and time may be specified instead of the initial visit date and time, and the appearance probability of the confirmed disease name for each predetermined period may be calculated based on this.

  Based on the result obtained by the disease name appearance distribution calculation means 4, the synonymous disease name determination means 5 obtains the disease name distances for all of the two pairs of disease names across the hospital departments, and determines whether the disease name has the same meaning. Determine whether or not. However, the two-term set whose disease name notation completely matches and the two-term set that can be identified as a known synonym disease name by referring to the synonym disease name holding means 2 are excluded from the determination. .

Here, if the appearance probability of the disease name n in i months is P (n, i), the disease name distance D (X, Y) between the disease name X and the disease name Y is calculated by the following (formula 1). .

And the disease name L and disease name M which satisfy | fill the following (Formula 2) over a department are determined to be synonymous disease names. However, (alpha) in a type | formula is a predetermined threshold value (positive real number). That is, when the disease name distance of the disease name M with respect to the disease name L is sufficiently smaller than the disease name distance of the other disease name j (for example, when the threshold value α = 5, compared to the other disease name distances) If it is smaller than 1/5, it is determined that the disease name L and the disease name M are synonymous, and these are selected as synonymous disease names.

  The synonymous disease name output means 6 outputs the synonymous disease name set selected by the synonymous disease name determination means 5. For example, the synonymous disease name is displayed on a display screen or printed on a medium such as paper so that the user is informed of the synonymous disease name. For example, the synonymous disease name is output to the synonymous disease name dictionary holding means 2 Then, it is excluded from the determination target by the subsequent synonymous disease name determination means 5.

Next, the synonymous disease name selection process according to this example will be described with reference to the process flowchart of FIG.
First, the disease name extraction unit 3 extracts a disease name from each medical document stored in the document storage unit 1 (step S1). In this example, a hospital department ID, a patient ID, and a document creation date are also extracted, and diagnostic data combining these is generated for each medical document.
Next, the disease name appearance distribution calculation means 4 arranges diagnosis data related to the same patient in time series (in order of document creation date / time) for diagnosis data related to a certain hospital department (step S2), and confirms the patient. The disease name and the date and time of first visit are specified (steps S3 to S5). When the identification of the confirmed disease name and the date and time of first visit is completed for all the patients related to the hospital department (step S6), the appearance probability for every predetermined period of the all confirmed disease names related to the hospital department is calculated (step S7). When the above processing is also performed for other hospital departments and the calculation of the appearance probabilities of all confirmed disease names in all hospital departments for a predetermined period is completed (step S8), the synonymous disease name determination means 5 determines the confirmed diseases of different hospital departments. A disease name distance is calculated for a combination of names (step S9), and based on this disease name distance, a different notation disease name expressing a synonymous disease is selected (step S10).

FIG. 4 illustrates the monthly appearance probability of the disease name calculated by the disease name appearance distribution calculating means 4 in each hospital department.
According to the figure, for example, in the internal medicine A of a certain hospital, the disease name “A / H1 subtype influenza” appears with a probability of “12.2%” in February 2007, and for example, another hospital In the internal medicine B, the disease name “A Hong Kong influenza” appears in March 2007 with a probability of “11.5%”.

  FIG. 5 shows the disease name distance calculated by the above-described (Equation 1) based on the monthly appearance probability of the disease name in each hospital department shown in FIG. That is, the disease name distance is calculated for all combinations of the disease name in internal medicine A and the disease name in internal medicine B. For example, “A / H1 subtype influenza” in internal medicine A and “A Hong Kong type influenza in internal medicine B” The disease name distance of “A / H1 subtype influenza” of internal medicine A and “A Soviet influenza” of internal medicine B is “0.0453”. .

  Here, when the above-described calculation of (Equation 2) is performed with the threshold α = 5, the synonymous disease name pair shown in FIG. 6 can be obtained. That is, for example, the disease name distance between “A / H1 subtype influenza” of internal medicine A and “A Soviet influenza” of internal medicine B is the name of another disease in internal medicine “A / H1 subtype influenza” and internal medicine B Because it is smaller than 1/5 of the disease name distance and it can be judged that the tendency of the temporal change of appearance probability is similar, “A / H1 subtype influenza” and “A Soviet type influenza” are selected as synonymous disease name pairs The

  In the description so far, synonymous disease names are compared by comparing temporal changes in the appearance probability of disease names (the ratio of the appearance frequency of the target disease name to the total appearance frequency of all disease names appearing during the same period). For example, synonymous diseases are compared by comparing temporal changes in the rate of change in the appearance frequency of disease names (the rate of increase or decrease in the appearance frequency of disease names in the target period relative to the appearance frequency of target disease names in the previous period). A name may be selected. In short, disease names having similar trends in the temporal change in the appearance frequency of disease names may be selected as synonymous disease names.

  Moreover, although it is determining whether it is a synonymous disease name for the two-tuples of disease names that cross hospital departments, in addition to the aspect of determining synonyms between disease names related to such different hospital departments, different doctors It may be an aspect in which the disease names related to each other are synonymously determined, and the disease names related to a common diagnostic entity (such as a hospital department or section or a doctor) considered to use a single disease name are excluded from determination And it is sufficient.

  The functional block diagram of the synonymous disease name selection device shown in FIG. 7 is an extension of the function of the synonymous disease name selection device shown in FIG. 1, and further includes an inter-hospital distance holding means 7 and a process end determination means 8. Yes.

The inter-hospital distance holding means 7 holds distance information related to the distance between hospitals, and the synonymous disease name determination means 5 refers to the distance information to acquire the distance between the hospitals, Synonymous judgment of disease names is performed in order from the combination. In addition, the coordinate value (latitude, longitude) of each hospital may be stored as distance information, and the distance between hospitals may be calculated based on the distance information. The distance between hospitals may be calculated in advance, and the distance information may be calculated in advance. You may make it hold | maintain as.
The process end determination means 8 determines to end the process when the distance between the closest hospitals among the unprocessed hospital combinations exceeds a predetermined threshold.

  According to the synonymous disease name selection apparatus, first, synonymous disease names are selected between the departments of two hospitals that are geographically closest to each other, and the obtained synonymous disease names are added to the synonymous disease name dictionary holding means 2. Next, synonymous disease names are selected between the departments of the two hospitals that are geographically closest to each other, and the obtained synonymous disease names are added to the synonymous disease name dictionary holding means 2. This process is repeated until an end determination is made by the process end determination means 8. In this way, by performing processing sequentially from geographically close hospitals, it is possible to carry out selection of synonymous disease names nationwide with high accuracy.

FIG. 8 shows a main hardware configuration of the synonymous disease name selection device of this example.
That is, the synonymous disease name selection apparatus of this example includes a CPU that performs various arithmetic processes, a RAM that is a work area of the CPU, a ROM that stores basic control programs, a program for realizing each function according to the present invention, and the like. Communicating with other devices such as an HDD that stores information, a liquid crystal display that displays and outputs information to the user, an input / output I / F that is an interface with devices such as a mouse and a keyboard that accept input of information from the user It is comprised with the computer which has hardware resources, such as communication I / F which is an interface which performs.

  Then, the program according to the present invention is read from the HDD, expanded in the RAM, and executed by the CPU, whereby the first storage means (document storage means 1) and the second storage means (synonymous disease) according to the present invention are executed. Name dictionary holding means 2), acquisition means (disease name extraction means 3 and disease name appearance distribution calculation means 4), determination means (synonymous disease name determination means 5), output means (synonymous disease name output means 6) and the like are synonymous. It is realized in the computer of the disease name selection device.

The program according to the present invention is provided to the practitioner of the present invention, for example, in a format for distributing an external storage medium such as a CD-ROM storing the program or a format for distributing via a network.
Moreover, each functional means of the synonymous disease name selection apparatus according to the present invention is not limited to the mode realized by the software configuration as in this example, and may be configured by a dedicated hardware module.
Moreover, each functional means of the synonymous disease name selection apparatus according to the present invention is not limited to an embodiment provided in one computer as in this example, and may be provided distributed in a plurality of computers.

It is a functional block diagram of the synonymous disease name selection apparatus which concerns on one Embodiment of this invention. It is a figure which illustrates synonymous disease name information which concerns on one Embodiment of this invention. It is a processing flowchart of the synonymous disease name selection process which concerns on one Embodiment of this invention. It is a figure which illustrates the appearance probability of the disease name which concerns on one Embodiment of this invention. It is a figure which illustrates the disease name distance which concerns on one Embodiment of this invention. It is a figure which illustrates the result of the synonymous disease name selection process which concerns on one Embodiment of this invention. It is a functional block diagram of the synonymous disease name selection apparatus which concerns on one Embodiment of this invention. It is a hardware block diagram of the synonymous disease name selection apparatus which concerns on one Embodiment of this invention.

Explanation of symbols

1: Document storage means,
2: Synonymous disease name dictionary holding means,
3: Disease name extraction means,
4: Disease name appearance distribution calculating means,
5: Synonymous disease name determination means,
6: Synonymous disease name output means,
7: Inter-hospital distance holding means,
8: Processing end determination means

Claims (6)

First storage means for storing diagnostic information including a patient's disease name and date and time related to the diagnosis;
An acquisition means for acquiring an appearance frequency for each predetermined period of each disease name based on a plurality of pieces of diagnostic information stored in the first storage means;
A determination means for determining a similarity between a tendency of a temporal change in appearance frequency relating to one disease name and a tendency of a temporal change in appearance frequency relating to another disease name;
An output means for selecting and outputting the disease names determined to have similar trends in the frequency of appearance frequency as synonyms;
A synonymous disease name selection device comprising: The synonymous disease name selection device further includes second storage means for storing synonymous disease name information in which synonymous disease names are associated with each other in advance.
The synonymous disease name selection apparatus according to claim 1, wherein the determination unit excludes disease names associated with each other by synonymous disease name information from being determined.   The synonymous disease name selection apparatus according to claim 2, wherein the output means outputs synonymous disease names to the second storage means and stores them as synonymous disease name information. The diagnostic information includes identification information of a diagnostic subject related to a patient diagnosis,
The synonymous disease name selection device according to any one of claims 1 to 3, wherein the determination unit excludes disease names related to a common diagnosis subject from being determined. The diagnostic information includes identification information of the hospital that diagnosed the patient,
The synonymous disease name selection device further includes third storage means for storing distance information related to the distance between hospitals,
The synonymous disease name selection apparatus according to any one of claims 1 to 4, wherein the determination unit performs the determination in order from a combination of hospitals having a short distance. Computer
First storage means for storing diagnostic information including a patient's disease name and date and time related to the diagnosis;
An acquisition means for acquiring an appearance frequency for each predetermined period of each disease name based on a plurality of pieces of diagnostic information stored in the first storage means;
A determination means for determining a similarity between a tendency of a temporal change in appearance frequency relating to one disease name and a tendency of a temporal change in appearance frequency relating to another disease name;
A synonym disease name selection program for functioning as an output means for selecting and outputting disease names determined to have similar trends in temporal changes in appearance frequency as synonyms.

Images