JP3915443B2 - Receipt processing system - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a technology that makes a character string converge to a standard character string and facilitates encoding, particularly when an insurer or an examination / payment institution is a medical institution. The present invention relates to a receipt processing system that performs statistical processing by encoding the names of wounds and diseases described in issued medical remuneration specifications.
[0002]
[Prior art]
Insurers such as health insurance associations are injured and sick in the medical remuneration statement (hereinafter referred to as “receipt”) issued by each medical institution for statistical processing of disease trends for the purpose of effective insurance business development, etc. Aggregation is performed using the disease names listed in the column.
Conventionally, these work has been done by workers with knowledge and experience reading the names of injuries and illnesses, and manually converting and counting the corresponding injuries and sickness codes. Since millions of receipts have arrived, it is required to be mechanized with a heavy burden in terms of time and cost for monthly aggregation.
In order to realize the mechanization of reading of the wound name by the conventional technique, for example, a character string obtained by character recognition of the wound name column by an OCR device (character reading device), and a record in which the standard wound name and the wound code are accumulated. There is a method of coding a corresponding disease name by collating a disease name master as a medium. However, in the name column of the injury and illness in the reception, supplementary items are added for more accurate description by doctors and medical institutions, or the names of diseases that are not standard disease names but are commonly used are used. Thus, it is not always easy to obtain the injury / illness code by collating the name of the injury described in the receipt with the injury / illness name master. For this reason, it is necessary to provide a user registration dictionary separately from the disease name master, and to register the injury / illness code corresponding to the injury / illness name that is not detected.
However, each time an unknown name described in a receipt is found, it is a great effort to register it in the user dictionary, and some of the character strings to which information that is not necessarily important for the user is added are included. There are many cases that do not exactly match. Therefore, various notations in wound names are automatically converted to standard notation, and when an unknown receipt description is detected, it is possible to know in advance whether the user needs to check visually. If it can be provided, work efficiency will be improved in each stage.
[0003]
[Problems to be solved by the invention]
In the conventional technique, in order to automatically code many names of injuries and diseases described in a reception, it is necessary to register more names of fluctuations and injuries in the user dictionary. Further, in the conventional technique, it is necessary to visually check all the names of the wounds that have not been detected at the time of matching, and add them to the user dictionary as necessary.
An object of the present invention is to convert a large number of written fluctuations into a standard disease name that can be easily compared with a wound disease name master with a small burden. Another object of the present invention is to analyze the contents and present the analysis result to the user even if there is no master disease name that is completely matched as a result of collation, so that the user can respond to the needs. The object is to provide an environment in which it is possible to select whether or not to modify.
[0004]
[Means for Solving the Problems]
In order to achieve the above object, the receipt processing system according to the present invention includes a character string conversion unit and a character string conversion master before collating a character string corresponding to a wound name described in the receipt with a wound name master. It is used to convert various notations that can occur on the character string of the name of the disease described in the receipt into a character string of the name of the disease name. This makes it easier to detect the name of the disease, and it is much more efficient than registering all the names of the disease that are included in the receipt in the user dictionary as in the conventional technology. Reduces the risk of Therefore, it is possible to reduce the number of names of sicknesses that the user has to confirm with sight. In addition, the receipt processing system of the present invention can present to the user how much the name of the wound described in the receipt matches the name of the master wound, and the name of the wound described in the receipt that does not completely match the name of the wound. A character string that is a factor that hinders matching is detected, and the contents of the character string are analyzed and presented to the user.
[0005]
DETAILED DESCRIPTION OF THE INVENTION
In the following, a receipt processing system according to an embodiment for supporting the implementation of the receipt description injury / illness coding process will be described.
Example 1
FIG. 1 is a basic configuration diagram of this embodiment. A computer apparatus 1 stores a program according to the present invention in a recording medium. The injured disease name data 2 supplied via an OCR device, punch input, external recording medium such as a floppy disk or a network connection with another computer device, etc., not shown, is a result of processing in the computer device 1. It is stored and accumulated in the receipt information database 3 in the form of an appropriate wound code. Alternatively, the processing result may be output on a form by a printer.
The computer apparatus 1 includes a character string conversion means 4, a wound disease name master collation means 7, a wound disease name master preprocessing means 8, a wound disease name generation means 10, a modifier master collation means 11, a collation result evaluation correction means 14, a character string conversion as a program. The master editing means 15 is composed of a character string conversion master 5, a collation table 6, a wound name master 9, a modifier master 12 and a collation result storage area 13 as a recording area, and a peripheral for the operator to manage and correct the processing result As a device, a display unit 16 such as a display and an input unit 17 such as a keyboard and a mouse are provided.
The general operation of this embodiment will be described with reference to FIG. The receipt description wound name obtained from the receipt description disease name data 2 is first searched by the character string conversion means 4 for the conversion target character string stored in the character string conversion master 5, and if detected, the converted character string. Is converted to At this time, the conversion target is not limited to the entire character string of the wound description disease name, but may be a part of the character string.
For example, the phrase “skin” is sometimes referred to as “skin” on the receipt. Therefore, if it is defined on the character string conversion master 5 that the character string “skin” to be converted is replaced with the character string “skin” after conversion, the name of the wound described in the receipt “contact skin fever” is “ It can be converted to a master injury name “contact dermatitis”. The name of the injury described in the receipt is converted by the character string conversion means 4 and then stored in the collation table 6.
Next, the master injury / sickness name stored in the injury / sickness name master 9 is searched by the injury / sickness name master verification means 7, and if detected, the master injury / sickness name detected in the verification table 6 (hereinafter referred to as the detected injury / sickness name), and The injury code is stored. When collating the receipt written disease name and the master wound name in the wound name master matching means 7, it is necessary not only to collate a part of the master wound name as it is, but also to convert it after converting it to another character string. In the wound name master pre-processing means 8, these character strings are generated and collated in the same manner.
For example, in the receipt computer processing system wound name master that can be used for the wound name master 9, the wound name “hypertension (symptom)” is “0901” of the middle classification (hereinafter referred to as the middle classification) of the disease classification for the social insurance chart of the Ministry of Health and Welfare. It is defined as hitting. This means that the name of the disease “high blood pressure” falls under the middle classification “0901”, whether or not “symptom” is subsequently described. Since the actual character string of injury and disease name described in the receipt can include any of “high blood pressure”, “hypertension”, and “hypertension (symptom)”, the wound disease name master pre-processing means 8 generates these three types of character strings. Then, each character string is collated with the name of the injury described in the receipt.
As described above, the detected wound names are stored in the collation table 6. At this time, there may be a plurality of detected wound names. In that case, only the necessary wound names are extracted by the wound name generation means 10, and this is extracted. It is stored in the collation table 6 as a generated injury / illness name. If there is a single detected disease name, it is stored in the verification table 6 as a generated disease name. If a plurality of generated wound names are generated, or if they are not completely the same character string as the receipt-written wound name, the generated wound name string is removed from the received wound name string to generate a remaining character string. To store. At the same time, a parameter called a detection attribute is stored in the collation table 6 based on the wound name generation result. The detection attribute is a parameter that indicates the degree of coincidence between the name of the wound and the disease described in the receipt, and the type is “complete match” if the name of the wound is the same as the name of the wound and the name of the wound. When there is a column, “partial match”, “multiple detection” when there are a plurality of generated wound names, and “not detected” when no wound names are detected. As an example, the operation of the wound disease name master collating means 7 and the wound disease name generating means 10 when the name of the wound description disease “chronic suppurative otitis media acute exacerbation” is input is shown. In addition to “chronic suppurative otitis media acute exacerbation”, the wound name master verification means 7 detects “chronic suppurative otitis media”, “suppurative otitis media”, “middle otitis”, “otitis”, etc. In the process, “chronic suppurative otitis media acute exacerbation” is used as the name of the generated wound and stored in the matching table 6 together with the detection attribute “complete match”, but there is no remaining character string. . As another example, it is assumed that when a receipt-written wound name “face eczema” is input, only the master wound name “eczema” is detected by the wound name master verification unit 7. Therefore, the detected wound name and the generated wound name are both “eczema” and “face” as the remaining character string, and the detection attribute is “partial match”.
When the remaining character string is generated by the injury / illness name generation means 10, the modifier master collating means 11 searches the remaining character string for a modifier phrase stored in the modifier word master 12. When a modifier is detected, the modifier and the corresponding modifier attribute are stored in the collation table 6. The modifier attribute is a parameter indicating the type of phrase included in addition to the name of the wound, and will be used later to evaluate the coding result.
For example, in the case of the name of the wound described in the receipt of “left forearm” contact dermatitis, the generated wound name generated by the wound name generator 10 is “contact dermatitis”, and the remaining character string is “left forearm”. The modifier master collating means 11 detects the modifier and its modifier attribute as “left”: “position”, “forearm”: “part”, “)”: “symbol etc.”, respectively. Details of the modification attribute will be described later.
The respective data stored in the collation table 6 by the above steps are accumulated in the collation result storage area 13.
The user refers to and corrects arbitrary data in the collation result storage area 13 or data in the collation table 6 via the collation result evaluation and correction means 14. The collation result evaluation / correction means 14 has a function of setting data conditions to be referred to by the user in advance, and detection attributes and modification attributes are used as conditional expressions. The collation result evaluation and correction means 14 displays the data satisfying the set conditions on the display device connected to the computer device 1, and the user compares and confirms the name of the wound described in the receipt and the name of the generated wound and disease based on the information, It can be directly selected from the disease name master 9 as required. At the same time, the character string conversion master is edited by the character string conversion master editing means 15 and the conversion target character string and the converted character string are added, so that the same correction can be omitted from the next time.
The collation result evaluation / correction means 14 can refer to the collation result storage area or the data of the collation table 6. This is performed by collating all of the receipt written disease name data 2 once in batch processing and then storing the collation result. When editing data belonging to a specific detection attribute or modification attribute stored in the area 13, or when checking and correcting at any time during the code output process with reference to the collation table 6, the user can change the data according to the operation mode. This is so that the result of the wound code output can be confirmed and corrected by this method.
As described above, the data corrected by the user as needed, that is, the name of the wound and the name of the wound and the name of the wound and the wound code are output as final results and stored in the receipt information database 3 or printed on a form. Is done.
Further, by using the character string conversion master editing means 15, the same name written in the next receipt is recorded by registering the name of the wound described in the receipt and the name of the generated disease in the character string conversion master 5 as a character string before conversion and a character string after conversion. Can be processed as a detection attribute “complete match”, and it is possible to provide a uniform disease and disease coding process without bothering the user.
FIG. 3 is a diagram showing a specific example of the character string conversion master 5 of the present embodiment. The character string conversion master 5 stores a conversion target character string and a corresponding converted character string. Since the character string conversion master 5 is installed for the purpose of converging various notation fluctuations of the wound name described in the receipt into the notation of the master wound name described in the wound name master 9, the character to be converted to be stored is stored. Columns and converted character strings, such as different notation commonly found in the name of the wound description in the receipt, such as differences in Hiragana / Katakana / Kanji notation ("Biran", "Bilan", "糜爛"), "Neck"), presence / absence of long sound symbols ("Qatar" and "Qatar"), etc. are registered. In the character string conversion means 4, only the target character string portion of the wound description disease name character string is converted and collated with the disease name master 9. Therefore, unlike the collation method using the user dictionary described above, there is no need to register all the names of sicknesses and diseases that contain different character strings in the user dictionary, and only the character string conversion master for the different character string part is required. The same effect can be obtained simply by registering in 5. Therefore, maintenance of the master by the user becomes easy and the risk of registration omission can be greatly reduced.
FIG. 4 is a diagram showing a specific example of the disease name master 9 of the present embodiment. The wound name / master 9 stores the name of the wound and the corresponding wound code, but in this embodiment, the ICD-10 code is a code composed of English letters and 3-digit numbers, and a 4-digit number. The medium classification code is written together. Of course, in the wound name master 9, other wound codes may be used, or one or more wound codes may be written together.
FIG. 5 is a diagram showing a specific example of the modifier master 12 of the present embodiment. The modifier master 12 stores modifiers and modifier attributes to which the modifiers belong. The modifier attributes include symbols such as parentheses and “and” and “symbols” including conjunctions, “right” and “peripheral”. "Position" indicating the position in the body such as "Part" indicating the organ of the body such as "Forearm" or "Forehead", "Etiology" indicating the etiology such as "Allergy", Injury such as "Acute""Suffix" after the wound name such as "Progressive expression" indicating the course of the disease, "State expression" indicating the state of the wound such as "Multiple", "Patient attribution" indicating the patient attribute such as "Infant", "Sequence""and so on.
The modifier master 12 examines what character string remains when the master wound name detected from the wound name master 9 is not exactly the same as the name of the wound described in the receipt, and determines the importance of the mismatched portion. It is installed for the purpose of checking. For example, if the remaining character string is a symbol or “position”, it is considered that there is no significant meaning in medical insurance. The same may be considered when there is a “part” in the remaining character string. However, if the remaining character string is a “state expression”, “suffix”, or a character string that is not stored in the modifier master 12, it is better to check it with the eyes of a knowledgeable user just in case. . By using the modifier master 12 in this way, it is possible to achieve the significance of the reason for the mismatch when the name of the wound described in the receipt does not completely match the name of the wound, and the user visually confirms the operation by using the modifier. It becomes possible to control the number of injuries and diseases described in the reception. If there is no function as in the embodiment of the present invention, if the user does not visually check all the names of the received injuries and diseases that do not completely match the injured names of the masters, there is a coding process in which a large error is made. If there is, you will miss it. FIG. 6 is a diagram showing a specific example of the collation table 6 of the present embodiment. The collation table 6 can be divided into five areas 61 to 65.
The region 61 is the name of the wound described in the receipt that has been pre-processed by the character string conversion means 4. In this figure, as an example, the name of the disease described in the receipt “left forearm) contact skin fever is expressed by the character string conversion means 4. It is assumed that it has been converted into “left forearm) contact dermatitis” and stored in the collation table 6.
The area 62 is a detected disease name detected by the disease name master collating means 7, a start position and an end position in the received wound disease name character string, and an injury disease code of the detected disease name. In this specific example, “dermatitis” (injury / sickness code: 1201) and “contact dermatitis” (injury / sickness code: 1202) are detected with respect to the name of the injury described in the receipt “left forearm” contact dermatitis, The start position and end position are as shown in the figure. In the Detected Sickness Name String column, the symptom name string of the received description other than the detected illness name is indicated by “○” so that the position of the detected illness name in the entire written symptom name string is easy to understand. This is not always necessary. Further, in this example, there are two detected wound names, but the number of cases increases or decreases depending on the name of the wound described in the receipt.
A region 63 is a generated wound name, a generated wound name string and a wound code extracted by the wound name generation means 10. In this specific example, “contact dermatitis” more consistent with the name of the wound described in the receipt is adopted as the name of the generated wound.
In this example, the number of names of wounds and diseases is one, but the number of cases is increased or decreased depending on the names of wounds and diseases described in the receipt.
A region 64 is a remaining character string extracted by the wound name generation unit 10 and a detection attribute.
In this specific example, the remaining character string that is the remainder obtained by removing the generated disease name character string from the receipt written disease name character string is “left forearm”. The detection attribute is the name of the generated wound stored in the area 63, and there is a remaining character string.
A region 65 is a list of modifiers detected from the remaining character string in the modifier master collating unit 11. In the case of the detection modifier, the detection modifier character string, and the area 65, the above-described modifier attribute is stored. There are three types of modifiers and their modifier attributes detected in this example: “left”: “position”, “forearm”: “part”, “)”: “symbols, etc.” The number of cases will increase or decrease depending on the name of the wound.
FIG. 7 is a diagram showing a specific example of the collation result storage area 13 of the present embodiment. The collation result storage area 13 is a recording area for storing various data generated in the collation table 6 for each name described in the receipt. The data stored in the collation result storage area 13 is later compiled by the user via the collation result evaluation and correction means 14. Are finally confirmed and corrected, and finally output to the receipt information database 3 or the like. The data stored in the collation result storage area 13 is the name of the wound described in the receipt from the left in FIG. 6, the detection attribute, the number of generated wounds and disease names, and the set of generated wound names and wound codes are the number of generated wounds and disease names, the number of modifiers, There are as many modifiers and modifier attributes as there are modifiers.
(Example 2)
In the second embodiment, detailed operation of each component of the receipt processing system of the present invention will be described.
FIG. 8 is a flowchart showing the detailed operation of the character string conversion means 4. First, in step 401, one character string for the name of the disease described in the receipt is fetched from the data 2 for the name of the disease described in the receipt. In step 402, one character string to be converted is fetched from the character string conversion master 5, and the character string to be converted is not included in the character string of the name of the disease described in step 401 in step 403 or the same character is included. Matches whether it is not a column.
If a character string to be converted is included in the character string of the name of the disease described in the receipt, the character string to be converted is replaced with the corresponding converted character string in step 404, and the character string conversion master 5 performs the next conversion. Read the target character string. The series of operations from step 402 to step 404 is repeated until it is confirmed in step 405 that all the conversion target character strings stored in the character string conversion master 5 have been checked against the receipt-written wound name character string.
Next, in step 406, the receipt-written disease name character string subjected to the conversion process is stored at a predetermined position in the collation table 6, and the operation of the character string conversion unit 4 is terminated.
Due to this operation processing, the different character string found in the name of the wound described in the receipt is converted into a character string having the same notation as the master wound name stored in the wound name master 9, so that it is detected by the wound name master matching means 7. The number of the name of the disease described in the receipt in which the name of the wound is generated increases.
FIG. 9 is a flowchart showing the detailed operation of the wound name master collating means 7.
In step 701, the character string converting process 4 takes in the character string for the name of the disease described in the receipt from the collation table 6. In step 702, one master wound name is fetched from the wound name master 9, and in step 703 the master wound name string is not included in the received wound name string or the same character string. Check if it is not.
If the master injury / disorder name character string is detected in the receipt description injury / disorder name character string, in step 707, the master injury / disorder name character string is from the first character to the first character of the receipt description injury / disorder name character string. Are matched, and values are obtained as a start position and an end position, respectively.
Thereafter, in step 708, the wound code corresponding to the master wound name is acquired from the wound name master 9, and in step 709, the detected wound name, detected wound name character string, wound code, start position, end The position is stored, and the operation of the wound name master collating means 7 ends. The detected wound name and the detected wound name character string here store the master wound name and character name string.
In step 703, if the master wound name string is not included in the received wound name string, it is not checked in step 704 whether the master wound name is the pre-processing target wound name string. Must not. The pre-treatment target master injury / sickness name here is a master injury / sickness name in which a part of the character string is enclosed in parentheses, as in the above-mentioned “high blood pressure (symptom)”. Any character string of “symptom” or “hypertension (symptom)” is acceptable. Therefore, in step 705, the parenthesis character is searched for () parenthesis characters, and if there is a character string written in both parentheses (), if there is no character string, three types of character strings are generated when there is no () parenthesis. These are used as collation character strings, and in step 706, as in step 703, it is collated whether they are not included in the receipt description disease name character string or are not the same character string.
Hereinafter, the operations in Step 708 and Step 709 are the same as those in the case of the master injury / illness name. However, at that time, the master wound name is stored in the detected wound name, but the detected matching character string is stored in the detected wound name character string. This is because, when comparing a plurality of wound names in the wound name generation means 10 to be described in detail later, there is a character string such as (symptoms) that is not included in the received wound name in the detected wound names. This is because it is difficult to compare as a column. A series of operations up to step 701 and step 709 are repeated until it is confirmed in step 710 that the matching between the name of the wound described in the receipt and all the names of the master wounds is completed. Note that the order of step 707 and step 708 may be reversed during a series of processing.
FIG. 10 is a flowchart showing the detailed operation of the wound name generation unit 10. In step 1001, the detected wound name, detected character string, wound code, start position, and end position are fetched from the collation table 6. If it is confirmed in step 1002 that there is only one detected disease name, it is checked in step 1005 whether the detected disease name or the detected disease name character string is the same as the name of the disease described in the receipt. The reason for examining both the detected wound name and the detected wound name character string is that, as described above, there is a possibility that the matching character string generated from the master wound name by the wound name master matching unit 7 is the same as the received wound name. Because there is.
If the name of the wound described in the receipt matches the name of the detected wound or the character string of the name of the detected wound, it means that the name of the wound described in the receipt is the master name of the wound. The detection attribute indicating the state of the detection result is “complete” in step 1009. The detected wound name is assigned to the generated wound name indicating the name of the master disease corresponding to the name of the wound described in the receipt. Although not shown in FIG. 10, the wound code of the detected wound name is also substituted as the wound code of the generated wound name.
If it is determined in step 1005 that the detected wound name or detected character string is not the same as the received wound name, the detected attribute is “partial match” in step 1008, and the detected wound name is substituted for the generated wound name. In this case, the detected disease name character string is assigned to the generated disease name character string.
On the other hand, if a plurality of detected disease names are confirmed in step 1002, one character string is determined in step 1003 based on a comparison evaluation method defined in advance from the start position and end position of each detected disease name character string. Is not included in the other character string, and the detected wound name included in the other character string is deleted.
Since some detected disease names are deleted through this process, it is checked in step 1004 how many detected disease names remain. When only one detected wound name remains, the processing from step 1005 is performed as in the case where there was originally only one detected wound name.
If there are a plurality of detected wound names in step 1004, the detected wound names are in an independent relationship. Therefore, in step 1007, the detection attribute is “multiple detection”, and the generated wound name and the generated wound name character string are A plurality of detected wound names and disease name strings are substituted.
If there is no detected disease name in step 1004, “undetected” is substituted for the detection attribute in step 1006, and nothing is substituted for the generated wound disease name or the generated wound name character string. In such a case, there may be a case in which there is no disease name in the name of the injury described in the receipt, or a case in which there is a disease name that is not detected in the disease name master 9 at all.
If the detection attribute is “complete match”, it means that the master wound name corresponding to the name of the wound described in the receipt has already been found. In step 1012, the generated wound name, wound code, and detection attribute are stored in the collation table 6. In step 1013, each data of the injury and illness table 6 is stored at a predetermined position in the verification result storage area 13.
In the case of other detection attributes, in step 1010, the wound description disease name character string is read from the collation table 6, a character string other than the character string detected as the generated disease name character string is checked, and this is used as the remaining character. Generate as a column. Therefore, when the detection attribute is “not detected”, the generated wound name string does not exist, and thus the received wound name string is generated as it is as the remaining character string.
In step 1011, the generated wound name, wound code, remaining character string, and detection attribute are stored at predetermined positions in the collation table 6.
FIG. 11 is a specific example of the comparative evaluation method used in the wound name generation means 10. In the figure, there are six positional relationships between the receipt-written disease name character string, the detected disease name character string A, and the detected disease name character string B. “○” means a letter, and “-” means nothing.
1) Included (both sides) is a character string that matches a part of both the detected wound name string A and the detected wound name string B with respect to the received wound name, but the detected wound name string A is more consistent The starting position starts early and continues until later. Therefore, the detected wound name character string A more correctly matches the name of the wound described in the receipt, and it is considered that this should be adopted. As well
2) Containment (front alignment), 3) Containment (backward alignment) is the case where only one of the start position and the end position is the same, and the detected wound name character string that is consistently long is adopted.
4) The deviation is a case where one detected disease name character string starts to coincide with the name of the disease described in the receipt earlier than the other, but the other coincides with the later and the relationship overlaps and deviates. In such a case, in the concrete plan of the present embodiment, the detected disease name character string that matches to the rear is adopted.
When the two detected disease name character strings do not overlap at all as in 5), both are adopted because they are independent of each other.
When the two detected wound name strings are exactly the same as in 6), it is considered that the same master disease name is registered twice in the wound name master 9, and either may be adopted. The name string A is adopted. These comparative evaluations are performed only with the detected injury name character string. If this is performed with the name of the detected disease, for example, if the disease name “hypertension” described in the receipt is detected in the matching character string “hypertension” of the master injury disease name “hypertension (symptom)”, the disease name described in the receipt is originally “(symptom ) ”Is not included, the detected injury / sickness name“ hypertension (symptom) ”does not hold as a comparative evaluation.
Such comparative evaluation is repeated between any two detected wound names until the remaining detected wound names become one or even if there are a plurality of detected wound names.
FIG. 12 is a flowchart showing the detailed operation of the modifier master collating means 11. In step 1101, the remaining character string is read from the collation table 6. At this time, as shown in FIG. 9, the name and character string written in the receipt of which the detection attribute is “complete match” is not applied because the modifier master collating unit 11 is not adapted. In step 1102, one modifier phrase is read from the modifier master 12, and in step 1103 it is checked against the remaining character string. If a modifier is included in a part of the remaining character string, or if the remaining character string and the modifier are the same character string, the modifier is read from the modifier master 12 as a detected modifier in step 1104. The modifier attribute corresponding to the detected modifier is stored in the collation table 6.
A series of operations from step 1102 to step 1104 is repeated until it is confirmed in step 1105 that the remaining character string and all modifiers are referred to. Thereafter, in step 1106, all of the data on the collation table 6 that has a predetermined position stored in the collation result storage area 13 is passed to the collation result storage area 13.
FIG. 13 is a flowchart showing detailed operations of the collation result evaluation correcting means 14 and the character string conversion master editing means 15. In the process so far, the type of the master injury / sickness name that matches the name of the injury / sickness / sickness of the read receipt, its injury / sickness code, the detection attribute, and the matching attribute, what type does not match completely due to the modification attribute? Whether or not the qualifier is preventing complete matching with the name of the master injury or illness is stored in the collation table 6.
Therefore, if the user selects the detection attribute and modification attribute that he / she wants to visually check in advance, and can browse only the name of the wound description and related information belonging to those attributes, the necessary visual confirmation can be performed very efficiently. It becomes possible. Moreover, if the conversion error is noticed and corrected at the time of visual confirmation, and the information is registered in the character string conversion master 5 at the same time, the same error will not be made from the next time. The collation result evaluation correcting means 14 and the character string conversion master editing means 15 are installed for these purposes.
In step 1401, first, from the collation table 6 or the collation result storage area 13, the detection attribute and the descriptive name of the written description of the disease belonging to the receipt attribute belonging to the modified attribute, the generated injured disease name, the injured disease code, the modifier, the modified attribute Is displayed on a display device or the like connected to the computer device 1. What can be displayed from both the collation table 6 and the collation result storage area 13 is the collation result storage area 13 and the code output when editing all the recorded injuries and sickness names once in batch processing and then editing from the accumulated data This is because it is more efficient to display from the collation table 6 to check and correct at any time during processing.
If the user visually confirms the name of the wound described in step 1402 and the related information and determines that there is no problem in the output of the wound code, the verification table 6 or the verification result storage area 13 is checked in step 1408. Data is output to the receipt information database 3 or the like. The name of the injury described in the receipt and the related information of the attribute that the user did not select to display on the display device are also considered to have been determined to have no problem, and are output to the receipt information database 3 or the like. If it is determined in step 1402 that there is a problem, the user can select the correct master injury / illness name using a screen in which the injury / patient name master can be searched in step 1403. The master wound name and wound code are replaced with the generated wound name in step 1404 and stored in the matching table 6 or the matching result storage area 13.
Here, the user can additionally register the corrected result in the character string conversion master 5 so as not to repeatedly correct it in the future. If registration is selected in step 1405, the receipt written wound name character string is additionally registered in the character string conversion master as the conversion target character string, and the selected master wound name is converted into a character string, and the process proceeds to step 1408. If the user does not register the entire character string for the name of the disease described in the receipt, but wants to register a part of the character string, it is possible to select editing of the character string conversion master in step 1406. In that case, a screen on which a conversion target character string to be newly added in step 1502 and a converted character string can be input is displayed. Next, in step 1503, the character string to be converted is checked against all the master injury and disease names in the injury and disease name master 9. This means that if a newly added string conversion master item is set so that the master injury / illness name is rewritten, inconsistency will occur as a string conversion master item that functions to make it easier to match the original injury / illness name. It is because it ends up. Therefore, in the event that the conversion target character string is detected from the master injury / illness name in the injury / patient name master 9 in step 1504, it is determined in step 1505 that this conversion object character string cannot be registered in the character string conversion master 5. After the display, the process returns to step 1401.
If the conversion target character string is not detected from the master injury / illness name of the injury / patient name master 9 in step 1504, the conversion target character string and the converted character string are additionally registered in the character string conversion master 5 in step 1407, and step 1401 The character string for the name of the disease described in the receipt is stored in the collation table, and the character string conversion is performed again. If the problem is solved when returning to the collation result evaluation / correction means 14 again, it should be output to the receipt information database in step 1408 without being displayed.
The above explanation is based on the assumption that the user replaces the name of the name of the disease with a problem with the correct master name, but if the detection attribute is actually set to “not detected” as the display target, If the name of the wound described in the receipt is not actually the name of the wound, it may be possible to discard the name of the wound described in the receipt. In addition, if the detection attribute is set to “multiple matches” as the display target, the wound description should originally contain only one wound name, so either one is selected or a new wound name master Although it should be selected from 9, it is necessary to be able to select a plurality of names if it is considered that a plurality of names of injuries and diseases are included and taken as the names of injuries and diseases described in the receipt, such as “hypertension and hyperlipidemia”.
14 and 15 show specific examples of display screens in the collation result evaluation correcting means 14 and the character string conversion master editing means 15.
The user refers to the injury / illness coding processing results accumulated in the matching result storage area 13 using the table screen shown in FIG. 14, and extracts and browses only the matching results belonging to the detection attribute and the modification attribute to be visually confirmed. can do. In this specific example, the name of the wound described in the receipt, the detection attribute, the modifier / modification attribute can be displayed in F1, and the user selects the detection attribute and the modification attribute to be visually confirmed by selecting the check boxes of F2 and F3, respectively. Only matching results that meet the conditions can be displayed in F1. If the user checks the correction column of F1 for the collation result to be confirmed and corrected and presses the F4 button, a correction / editing display screen on which more detailed information regarding each collation result is displayed is displayed.
FIG. 15 shows an example of the correction / edit display screen. In this specific example, the name of the wound described in the receipt selected in G1, the name of the generated wound and its disease code in G2, and the modification phrase and modification attribute detected from the remaining character string are displayed in G4. If the user compares these and determines that there is no problem with this conversion, the user can press the selection button of G3 and adopt the generated wound name. However, if it is determined that the name is not the generated injury / illness name, the intended injury / illness name or a part of the character string is input to G5 and G6 is pressed. It is retrieved from the wound name master 9 and displayed. The user can select from these master disease names and add them to the G8 corrected disease name column with the G9 button, or delete the master disease names once added with the G10 button and edit them.
After the corrected injury / illness name is confirmed, the content is reflected when the G13 button is pressed, but when the G11 button is pressed, it is also reflected in the character string conversion master at the same time. In this case, the entire name written on the receipt will be registered as the conversion target character string, and the corrected wound name will be registered as the corrected character string. However, as described above, only a part of the character string is registered in the character string conversion master. To call up the character string conversion master edit screen, press the G12 button.
For example, the name of the wound name is red, the modifier attribute “symbol / other” is gray, the “position” is If it is displayed in a color such as light blue, it is possible to understand a portion to be noticed at a glance, and to promptly make an accurate judgment sensuously. In the present embodiment, the correction result can be arbitrarily registered in the character string conversion master 5 by the user. However, the corrected injured disease name and corrected character string are automatically registered in the character string conversion master. It can also be done. In this case, since the number of words registered in the character string conversion master 5 increases with time, for example, as shown in FIG. If they are integrated, an unnecessary increase in the character string conversion master can be avoided by deleting character string conversion words that are not frequently referenced. In addition, the deletion process may be automatically performed when the number of reference times is less than a certain number. Here, the reference count is stored, but instead, the previous reference date is stored and deleted if the period is too long, or the reference count per fixed period is stored and the value is too low. You can delete it.
As described above, the notation fluctuation of the written description on the receipt is converged to the notation character string of the master wound name stored in the wound name master 9 by the character string conversion means 4 and the character string conversion master 5. In the means 7, the name of the master injury / sickness is easily detected. In addition, since conversion is performed on a character string, notation fluctuations are removed for all the names described in the disease including the character string, and all the names described in the receptor are registered in the user dictionary. It is much more efficient than it, and there is less risk of omission of registration.
In addition, the detection attribute can indicate to the user how much the name of the wound description matches with the name of the master injury, and the name of the description described in the receipt that did not exactly match the name of the damage of the master Since it is possible to detect a modifier and present its modification attributes to the user, the user must visually check all the names of the wounds and diseases that have not been completely matched and the generated wound names and codes. Rather, it is possible to control the name of the disease described in the receipt that is visually confirmed according to the operation mode and the burden of confirmation work.
Note that the present invention is not only effective in converging fluctuations of written disease names in the receipt to character string representations stored in the disease name master, but is also effective when, for example, questionnaires having free description columns are automatically aggregated. .
[0006]
【The invention's effect】
As described above, the method for encoding the name of a disease described in the receipt of the present invention and the recording medium storing the code for outputting the code of the disease described in the receipt are generated by different notations as in the prior art when encoding the name of the disease described in the receipt. Significant effect of greatly improving work efficiency and reliability of code output because it takes time and effort to register all the fluctuations of notation of various injury and illness names in the user dictionary, and it is not necessary to perform operations that are likely to cause registration errors. There is.
Further, according to the present invention, all the names of wounds and diseases that have not been detected at the time of collation of the names of the names on the receipt in the prior art must be confirmed by human eyes and added to the user dictionary as necessary. Even if there is no master injury or illness name that matches the above, by analyzing the content and presenting it to the user, by providing an environment where the user can select whether to correct the output result according to needs, There is an effect that the optimum operation according to the user needs and scale becomes possible.
[Brief description of the drawings]
FIG. 1 is a basic configuration diagram showing an embodiment of the present invention.
FIG. 2 is a process flow diagram illustrating one embodiment of the present invention.
FIG. 3 is a diagram illustrating an example of a character string conversion master.
FIG. 4 is a diagram showing an example of a wound name master.
FIG. 5 is a diagram showing an example of a modifier master.
FIG. 6 is a diagram showing an example of a collation table.
FIG. 7 is a diagram showing an example of a collation result storage area.
FIG. 8 is a flowchart showing an example of character string conversion means.
FIG. 9 is a flowchart showing an example of a wound name master verification unit.
FIG. 10 is a flowchart showing an example of a wound name generation unit.
FIG. 11 is an explanatory diagram showing an example of a comparative evaluation method for a plurality of detected wound names.
FIG. 12 is a flowchart showing an example of a modifier master collating unit.
FIG. 13 is a flowchart showing an example of collation result evaluation correcting means and character string conversion master editing means.
FIG. 14 is a diagram of a search extraction screen example in the collation result evaluation correcting unit and the character string conversion master editing unit.
FIG. 15 is a diagram showing an example of correction and editing screens in the collation result evaluation correcting means and the character string conversion master editing means.
FIG. 16 is a diagram showing another example of a character string conversion master.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... Computer apparatus, 2 ... Receipt wound disease name data, 3 ... Receipt information database, 4 ... Character string conversion means, 5 ... Character string conversion master, 6 ... Collation table, 7 ... Wound and disease name master collation means, 8 ... Wound and disease name Master pre-processing means, 9 ... Wound and illness name master, 10 ... Wound and illness name generation means, 11 ... Qualifier master collation means, 12 ... Qualifier master, 13 ... Collation result storage area, 14 ... Collation result evaluation correction means, 15 ... Character Column conversion master editing means, 16 ... display means, 17 ... input means.

Claims (7)

The wound description disease name database storing the injury name data described in the receipt,
A wound name and disease name master that stores a standard wound name and wound name and a wound code corresponding to the master wound name,
Obtains the name of the wound described in the receipt from the name database of the wound described in the receipt, collates the character string written in the receipt of the disease with the name of the wound name master, and matches the whole or a part of the character string described in the name of the disease described in the receipt Detecting a master wound name, the detected wound name, wound code, and the detected wound name and the start position and the end position of the detected wound name character string in the receipt are stored in a collation table. Master verification means;
Among the detected wound names extracted by the wound name master collating means, the wound name generating means for extracting the most applicable to the written description of the wound described in the receipt,
An output means for outputting the generated wound name and wound code extracted by the wound name generating means;
A receipt information database for accumulating the names of generated wounds and diseases output by the output means,
The wound name generation means
From the matching table, obtain the detected wound name, wound code, start position and end position,
Determining whether the acquired name of the detected disease is one or more,
In the case where there are a plurality of detected disease names, from the start position and end position of the acquired detected disease name character string, if one character string is in an implication relationship with the other character string, Delete the detected wound name
If there is no detected injury name, the detection attribute is not detected,
In the case where the number of remaining detected wound names is plural, the detection attribute is plural detection,
If the number of the detected wound names is one, if the detected wound name detected from the wound name master matching means matches the received wound name described in the receipt, the detection attribute is completely set. In addition, if the detected wound name detected from the wound name master collating means and the receipt described wound name described in the receipt do not match, the detection attribute is a partial match,
A receipt processing system, wherein a generated wound name, wound code, and detection attribute are stored in the collation table. The receipt processing system according to claim 1, further comprising: a modifier master storing a modifier and a modifier attribute to which the modifier belongs, and a modifier master collating unit;
The wound name generation unit obtains the receipt wound name string described in the detected wound name receipt from the received wound name string, and sets a character string other than the character string detected as the generated wound name string as a remaining character. As a column,
The modifier master collating unit obtains the remaining character string stored in the collation table, stores a modifier attribute corresponding to the detected modifier word detected from the modifier master in the inquiry table, and the output unit. The receipt processing system which outputs the modification attribute corresponding to the said detection modification phrase, and accumulate | stores in the said receipt information database.   The receipt processing system according to claim 1, further comprising: a character string conversion master in which a conversion target character string described in the receipt and a converted character string for converting the conversion target character string are registered, and the character string conversion The means collates all or part of the character string of the wound name described in the acquired receipt with the character string conversion master, converts it to the converted character string, and then collates with the disease name master. And a receipt processing system.   The receipt processing system according to claim 3, further comprising means for selecting a detection attribute and a modification attribute, wherein the wound name is a generated wound name extracted by the wound name generation unit and belongs to the selected attribute, and A receipt processing system, comprising: a matching result evaluation correcting unit that displays the related information together with a character string of a name of a disease described in a receipt described in a receipt, accepts the correction, and accumulates it in the receipt information database.   5. The receipt processing system according to claim 4, wherein the modification attribute is any one or more of a site, an etiology, a course expression, a state expression, patient attribution, and a suffix.   5. The receipt processing system according to claim 4, further comprising: a character string conversion master editing unit that registers the correction received by the collation result evaluation correction unit in the character string conversion master as a conversion target character string and a converted character string. A receipt processing system characterized by that.   2. The receipt processing system according to claim 1, wherein when there are a plurality of detected disease names, the wound name generation unit performs a plurality of detections on a position of a character string of a detected wound name with respect to a character string of the wound disease name described in the receipt. In the case of the inclusion relationship of both end positions by comparing the name of injury and illness and judging the inclusion relationship of both end positions, the inclusion relationship of front or rear alignment, deviation position, independent position, the same position Deletes the name of the detected wound or disease included, deletes the name of the shorter detected wound or disease in the case of the forward or backward alignment, and lists the names of the detected wounds or diseases in the case of independent positions. A receipt processing system in which any one or more is left as a detected wound name in the same position without being deleted.

Images