JP2004355633A - Device, method, and program for reading information of medical bill, and device, method, and program for supporting examination of medical bill - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To make it easy to find out errors of medical bills and to exactly distribute human resources in accordance with the degree of inaccuracy of each medical bill or the degree of concentration on the medical bill. <P>SOLUTION: Character strings having high possibility of being described in medical bills are prepared as a dictionary, and keywords agreeing with contents of the dictionary are registered as text data (S01) when they are found in combinations of component information read from medical bills as images. The keywords are extracted from items described in medical bills to perform data mining (S03). A rule having high reliability is extracted on the basis of rules obtained by past examination, and this rule is used to sort medical bills into a group of medical bills having few errors and a group of medical bills having a lot of errors. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a medical fee statement information reading apparatus and method, and a program thereof, a medical fee statement examination support apparatus and method, and a program thereof.

The insured of public medical insurance pays a partial co-payment to consult, treat, prescribe, and hospitalize in hospitals and clinics, pharmacies, dental clinics and other medical institutions (hereinafter referred to as medical institutions, etc.). And other medical services. Also, when a doctor's prescription is received, a prescription for a drug can be made at an insurance pharmacy by paying a partial contribution.
It is necessary for a medical institution or the like to request an insurer of public medical insurance to pay a contribution to be paid by the insurer as a medical fee.

Here, a request for payment of a medical fee to an insurer from a medical institution or the like is made via a bill called a medical fee statement (hereinafter, referred to as a receipt). The claim is
Matters concerning the contents of medical treatment and dispensing, such as the number of treatments and the type and number of medications for the insured, are described in detail, and payment is made in accordance with the contents of the description.

An examination / payment organization is interposed between the medical institution and the insurer as an organization for conducting a review of the receipt in order to confirm the adequacy of the receipt and to realize smooth payment of the medical fee.
Therefore, the claims made by medical institutions are collected by the examination and payment organization, and firstly the administrative assistants check for any doubts, followed by doctors with sufficient technical and technical knowledge (hereinafter referred to as “doctors”). , A reviewer). As a result, a claim with a basic error is returned to a medical institution or the like, and a claim that is appropriate or determined to be appropriate by examination is sent to each insurer. Conventionally, most of these processes have been performed by paper receipt.

  However, since most of the claims currently being handled in the medical fee billing affairs are paper, administrative assistants and reviewers must inspect and judge this huge paper claim one by one. Work is unscientific and inefficient. For this reason, despite the fact that the degree of error and the degree to which attention must be paid in response to them are quite different, the fact is that similar human resources are being allocated, and sufficient administrative assistance and The effect of the examination cannot be obtained.

  Conventionally, such a system or method relating to a claim or insurance procedure includes a method in which an insurer receives a claim and settles the claim or returns the claim to a medical institution (see Patent Document 1). The insurance card data for each insured person is stored in the database of the medical information storage and management company that keeps it. Confirmation of the insurance card at the time of consultation and billing to the insurer are performed through this management company, and There is one that reduces the burden (see Patent Literature 2), and one that checks whether the insured person has the qualification of the insured person to prevent loss of medical expenses for the insurer (see Patent Document 3). .

JP 2001-344337 A ([0012], FIG. 1) JP-A-2002-41649 ([0034], FIG. 7) JP-A-2002-109069 ([0020] to [0028], FIG. 1)

However, none of the conventional technologies merely solves the problems of securing efficient clerical work and reducing the processing burden on insurers and insured persons of public medical institutions, medical institutions, and the like. However, it does not attempt to solve the problems such as securing efficient office work and reducing the processing load at the examination and payment organization.
The present invention has been made in view of the problems of securing efficient office work and reducing the processing load in the examination and payment institution, and it is possible to sufficiently improve the efficiency of office assistant work and review the load of examination work. In this way, errors in each claim can be easily found, and human resources can be appropriately allocated to each claim according to the degree of inappropriateness and the level of focus. It is another object of the present invention to provide a medical fee statement information reading apparatus and method, a program therefor, a medical fee statement examination support apparatus and method, and a program therefor.

  In order to solve the above-mentioned problems, the present invention has been aimed at improving business efficiency by converting paper receipts into text data, in particular, mechanically processing a part which could not be completely converted into text data with the conventional technology and using it for examination. . That is, a character string (keyword) that is likely to be described in the receipt is prepared as a dictionary, and if there is a keyword that matches this dictionary in the combination of parts cut out from the imaged paper Registered as text data, enabling machine processing. In the past, since the work was performed manually by using information on paper, the obtained burden reduction effect was large.

Further, data mining processing is performed on the items of the receipt information converted into text. Using a large amount of receipt data, a highly credible sorting rule is extracted based on the know-how obtained from past examinations, and the claims are sorted using this rule. A certain claim is divided into a less-recept group. As a result, it is possible to extract a claim that does not need to be examined and perform an efficient examination.
Furthermore, if new know-how can be obtained from those that have been examined, the viewpoint of data mining can be changed and new rules can be generated, which will be useful information for subsequent examination support.
In addition, by applying a judgment rule composed of keywords extracted at the time of reading to a group of claims having many erroneous claims, it is possible to further narrow down claims having a high possibility of error, and to obtain various angles. Inspections from the public and prioritized examinations can be performed, and the work of the person in charge of examinations is streamlined.
As a result, it is possible to sufficiently improve the efficiency of work and reduce the burden of review processing, facilitate the discovery of errors for each claim, and respond to the degree of error and the level of focus for each claim. Human resources can be appropriately allocated.

  As described above, according to the present invention, it is possible to easily find an error in each medical remuneration statement by improving the efficiency of the administrative assistant work and reducing / reviewing the burden of the examination processing, and having each claim It is possible to appropriately allocate human resources in accordance with the degree of inaccuracy and the degree of focus.

FIG. 1 is a diagram cited for describing a model for realizing a medical fee statement information reading method and a medical fee statement examination support method in the embodiment of the present invention, and a relationship between the model and a user.
In FIG. 1, reference numeral 1 denotes a medical institution, reference numeral 2 denotes an examination and payment institution, and reference numeral 3 denotes an insurer. Here, a mechanism in which a remuneration (medical remuneration) is charged to an insurer in accordance with medical care performed on a patient, such as an insured person, is also described.
The medical fee is charged by the medical institution 1 such as a hospital or a pharmacy to the examination and payment institution 2, and the examination and payment institution 2 examines the medical fee from the medical institution 1 and appropriately evaluates the medical reimbursement request if necessary. After the assessment, the results are totaled for each insurer, and the medical treatment fee is charged to the insurer 3 in which the patient is enrolled as an insured. The insurer pays the medical fee to the examination and payment institution, conducts examination and inspection by the insurer, and requests the examination and payment institution for a reexamination.

The medical institution 1 requests the examination and payment institution 2 for a medical fee based on a receipt prepared for each patient and each month. Most of the claims provided by the medical institution 1 are based on paper handwritten or printed by a personal computer or the like, but recently, text-based or code-based magnetic recording media have also been found.
Therefore, in the present embodiment, it is considered that the information described in the paper-based receipt is converted into text data because the efficiency of business cannot be obtained with the paper-based one.

That is, the receipt includes information on a prefecture number, a medical institution code, a patient's date of birth, and the like, all of which can be read by the OCR. However, the description of the medical practice such as the injury / illness name column and the summary column may have small characters, and if there is an overlap with the ruled line, misreading will increase.
Therefore, the characters read by the OCR or the like are recognized as components, and a combination thereof is considered. At this time, keywords that are likely to be described in the receipt are extracted from medical insurance related materials and the like (S01), and held as a receipt dictionary. When it is found, it is recognized and stored as text (keyword information).

At the present stage, most of the claims submitted by medical institutions are made of paper (indicated as “handwritten claims” and “paper claims” in the figure), but claims are made on magnetic recording media such as flexible disks. As described above, there is also a receptor. Some of the claims for magnetic recording media are in text form, shown as "Print Recept" in the figure, or are partially coded, shown as "Magnetic Recept" in the figure. Can be considered.
Therefore, the paper-based receipt data (keyword information) obtained from the OCR and keyword extraction and the claim requested by the magnetic recording medium are coded (S02) and registered in the database. As a result, the examination / payment institution 2 can handle the requests received in various forms in the transitional period during the transition to computerization in the same form, thereby improving the operational efficiency.

Further, as described above, the examination payment institution 2 obtains the receipt data, performs an administrative check of the receipt on the computer, and then performs the examination support processing (S03). The examination support processing means that the receipts after the administrative check are sorted into a group of receipts with many errors in accordance with the sorting rules and a group of claims with few errors in accordance with the sorting rules.
Data mining finds highly credible rules based on the know-how gained from the results of the examinations so far, and sorts the claims based on these rules. The data mining process is to process various and large amounts of data by a statistical / mathematical method. Here, association extraction for searching a combination pattern of highly relevant data is used.

If the receipts are sorted using the rules and the know-how of the judges described above, the receipt group that has many erroneous claims and the reception group that has few erroneous claims can be sorted, so that the reception group that does not need to be examined is extracted. And reduce the burden on reviewers. Furthermore, by performing a more focused and focused review on a group of claims with many erroneous claims, an efficient review can be realized. Also, if new know-how can be obtained from those that have actually been examined, it is possible to change the viewpoint of data mining and generate new rules, which will be effective information for subsequent examination support .
Also, by applying a judgment rule composed of keywords extracted at the time of reading to a group of claims with many erroneous claims, it is possible to further narrow down claims with a high possibility of error, and to use various angles. Inspections from the public and prioritized examinations can be performed, and the work of the person in charge of examinations is streamlined.

  Further, if the above is executed as the examination support, when the claim substance examination is performed (S04), it is possible to extract a desired claim under various conditions. In addition, since past receipts are also accumulated as data, search and inspection can be performed easily even during inspection and inspection, and efficient examination can be performed. For example, search for disease names that often cause errors, such as those that people pay attention to when conducting examinations on paper, examinations that focus on high medical expenses, or only those that are difficult to make mistakes in prescriptions such as colds. It is also possible to extract it.

FIG. 2 is a block diagram showing the internal configuration of a receipt management system 4 in which the medical remuneration statement information reading device according to the present embodiment is mounted, in a functionally developed manner.
In FIG. 2, the medical remuneration statement information reading device includes an image file 201, a character reading unit 202, a text file 203, a virtual character reading file 204, a receipt data importing unit 205, a receipt dictionary 206, and a keyword file. 207, a recording medium file 208, a code conversion unit 209, a code table 210, a code file 211, and a receipt dictionary generation unit 212.

The image file 201 stores an image when a paper-based receipt is read by OCR. The character reading unit 202 has a function of reading information such as a basic information column of a receipt from the image file 201, and further has a function of extracting a character line candidate such as a summary column and cutting out a character string as a component. Information such as the basic information column is stored in the text file 203, and the parts extracted from the character line candidates are stored in the virtual character reading file 204.
The receipt data capturing unit 205 compares all combinations of character string components stored in the virtual character reading file 204 with a reception dictionary 206 in which a frequently used character string described in the receipt is described in advance, When there is a character string that matches the receipt dictionary 206, it has a function of extracting the character string as text data, and the extracted character string is stored in the keyword file 207.

Further, the code conversion unit 209 converts the text data captured by the character reading unit 202 and the receipt data capture unit 205 and the text data of the receipt information requested by the magnetic recording medium captured in the recording medium file 208. It has a function of coding in accordance with the code table 210 and storing it in the code file 211.
Note that the receipt dictionary generation unit 212 has a function of extracting keywords that are likely to be described in the receipt from medical insurance related materials and the like in advance and generating a file as the receipt dictionary 206.

FIG. 4 is a flowchart cited for explaining the operation of the medical remuneration statement information reading device according to the present embodiment, and specifically shows a processing procedure of the medical remuneration statement information reading program of the present embodiment.
Hereinafter, the operation of the medical remuneration statement information reading apparatus of the present embodiment will be described in detail with reference to the flowchart shown in FIG.

First, the examination and payment institution 2 accepts a paper-based receipt from the medical institution 1 and a receipt recorded on a recording medium (S50).
In the case of a paper-based receipt, characters of the receipt information described in the receipt are read (S51), but as shown in FIG. 7, the receipt includes a prefectural number, a medical institution code, and the patient's date of birth. And an OCR readable area A, and a summary field B having a description part of the medical treatment. The former is supplied to the code converter immediately after reading the character, while the latter takes the image and converts it into text-based information by performing the following processing.

Reading of the printed characters described in the summary column B will be briefly described below.
First, the character reading unit 202 extracts a character line candidate and a character cutout candidate from the input image stored in the image file 201 (S52), and stores them in the virtual character reading file 204 (S53). The character line candidates are the coordinates of the vertices at the top, bottom, left, and right of the circumscribed rectangle, but may be the top, bottom, left, and right coordinate values of the character cutout candidates.

Next, the receipt data acquisition unit 205 reads out the character string cutout data stored in the virtual character reading file 204 in any combination, and a frequently used character string described in the receipt is registered in the receipt dictionary 206 in advance. Is compared (S54). Here, when there is a character string that matches the receipt dictionary 206, the character string is taken into the keyword file 207 as text data (S55). Thereafter, code conversion is performed together with the read data of the OCR readable area A and the receipt data requested by the magnetic medium (S56), and stored in a code file (S57).
In this way, the character string taken from the input image is recognized as part information, and all combinations of the part information and the keyword that is likely to be described in the claim from medical insurance related materials and the like are registered. If there is a matching character string in the data of the dictionary 206, it can be extracted as text data. By reading only the necessary parts to realize the examination support for receipts, it is possible to efficiently obtain the information in the summary column, which was previously impossible to read completely by OCR and often read incorrectly. Can be extracted much more efficiently.

  Also, paper-based receipt data previously obtained from OCR and keyword extraction, receipt data (image file 201, text file 203, keyword file 207) extracted from a description column and the like, and receipt data requested by a recording medium ( The magnetic medium file 208) and the data obtained by encoding the respective text data are collectively managed by the receipt management system 4 as a receipt DB. As a result, the examination / payment institution 2 can handle the requests received in various forms in the transitional period during the transition to computerization in the same manner in the same system, thereby improving the operational efficiency.

FIG. 3 is a block diagram showing the internal configuration of the medical fee remuneration statement examination support device according to the present embodiment in a functionally developed manner.
In FIG. 3, the medical remuneration statement examination support device includes a receipt DB 213, a sorting rule selection unit 214, a receipt sorting unit 215, a sorting rule DB 216, a determination rule selection unit 217, a receipt extraction unit 218, and a determination rule. It comprises a DB 219, a rule / error receipt image display unit 220, a sorting condition fetching unit 221, a sorting rule updating unit 222, a judgment condition fetching unit 223, and a judgment rule updating unit 224.

  The sorting rule selection unit 214 has a function of displaying sorting rules from the sorting rule DB 216 and selecting a sorting rule to be applied from the displayed rules as needed.

  The receipt sorting unit 215 applies the rule selected by the sorting rule selection unit 210, and stores the receipt DB 213, which is collectively managed by the receipt management system 4, as a receipt group containing many erroneous claims and a erroneous claim. Has a function to sort into a small number of receipt groups.

  Further, the sorting condition fetching unit 221 has a function of determining a sorting rule from the evaluation of the correlation rule analysis using a large amount of past receipt data. As an application condition of this sorting rule, a support degree indicating the generality of the rule and a certainty degree indicating the reliability of the rule are used. For example, “support 2.0% or more, certainty 99.98% or more” To be determined. These rules are evaluated as needed from the result of newly performed data mining using the latest receipt data, and the sorting rule updating unit 222 updates the sorting rule DB 216.

  The determination rule selection unit 217 has a function of displaying a determination rule from the determination rule DB 219 and selecting a determination rule to be applied from the display as needed.

  The receipt extraction unit 218 has a function of extracting a receipt corresponding to the rule selected by the determination rule selection unit 217.

  Further, the determination condition capturing unit 223 inputs a determination rule determined from the results of past examinations and know-how. As these rules, rules generated from the result of the examination of the new receipt data are input, registered / corrected in the judgment rule update unit 224, and the judgment rule DB 219 is updated.

On the other hand, the rule / error receipt image display unit 216 displays a sorting rule, a judgment rule, and a receipt including an error extracted by applying the selected judgment rule.
Also, here, the item that is determined as a receipt containing an error is highlighted by a red line or the like, and the determination rule applied at this time is displayed in a sticky note. A message is displayed.

FIG. 5 and FIG. 6 are flowcharts cited for explaining the operation of the medical remuneration statement examination support device in the present embodiment, and specifically, the processing procedure of the medical remuneration statement examination support program of the present embodiment. Is shown.
Hereinafter, the operation of the medical fee statement examination support device of the present embodiment will be described in detail with reference to the flowcharts shown in FIGS.

Here, the sorting rule selection unit 214 selects a rule to be applied as necessary from rules obtained by analyzing correlation rules of past receipt data (S60). At this time, the contents of the sorting rule DB 216 are displayed on the rule / error receipt image display unit 220, and a rule to be applied can be selected. By pressing a “rule application” button on the screen (S 61 “ON”), the selected rule is applied, and the receipt sorting unit 215 receives a group of claims that are likely to include an error and a group of claims that are unlikely to include an error. They are sorted into groups (S63). The following determination rule is applied to a group of claims determined to have many errors containing errors.

The examiner extracts a claim from the group of claims that are likely to include the classified error by applying a further determination rule by the claim extraction unit 218 and performs a review.
First, a rule stored in the determination rule DB 219 is displayed, and a condition for selecting or extracting a determination rule is input (S64 “input”). By pressing the "apply rule" button (S65 "ON"), the selected rule is applied (S66), and only the recepts meeting the conditions are extracted from the recept group and displayed in a list (S67). When the user selects a desired reception from the list and presses the “image display” button (“ON” in S68), the reception can be displayed as an image (S69). At this time, places that violate the rules are marked.
In addition, the name of the determination rule applied at that time is displayed as a sticky note, and by clicking the sticky note ("YES" in S70), a message indicating the reason for the determination as a receipt containing an error is displayed (S71).

FIG. 8 shows an example of a description item of a receipt used when creating the sorting rule described above. Here, the month of medical treatment, number of medical treatment days, age, gender, etc. are shown in the basic information column, the number of disease names and disease classifications are shown in the injury / illustration column, the presence / absence of each medical treatment is given in the score column, and the presence / absence of a decision point is found in the determination column. Is read.

9, 10, and 11 show screen examples of the rule / error-reception image display unit 220 when applying the sorting rule and the determination rule.
FIG. 9 shows a selection screen for selecting a sorting rule, FIG. 10 shows a selection screen for selecting a determination rule, and FIG. 11 shows a screen on which an image of the extracted receipt is displayed.
In the screen for selecting a sorting rule shown in FIG. 9, the sorting rule read from the sorting rule DB 216 is displayed as the rule (a) used for sorting. This rule can be selected as needed.
The screen displayed when the determination rule of FIG. 10 is used displays a determination rule list (b) used in applying the rule and a keyword list (c) in the receipt extracted by the receipt data capturing unit 205. There are two buttons: "Apply rules" that is clicked by the examiner when applying the judgment rules to narrow down the claims, and "Screen display" that is executed by the examiner to output the narrowed receipts as images. Buttons are assigned. On the other hand, FIG. 11 shows a state after the user presses the “execute rule” button in FIG. 10 and extracts a receipt conforming to the selected rule and displays it on the screen. A list (d) of the rule application results of the received claim, an application rule name (e) used for narrowing down the claims displayed on the screen, and an image (f) of the narrowed claim are displayed.

As described above, according to the present invention, a character string (keyword) that is likely to be described in a receipt is prepared as a dictionary, and a keyword that matches this dictionary is included in a combination of component information read as an image from the receipt. If there is, register it as text data and read the summary column, which was not possible with OCR in the past. It has a function that can obtain important data for performing.
In addition, sorting rules are created based on the sorting criteria based on past screening results obtained by data mining and the know-how of the judges, and classified into the receiving group with many errors and the receiving group with few errors. . Further, by applying a judgment rule from a keyword to a group of claims having many errors, it is possible to extract a claim that requires a intensive examination and perform the examination.
Also, if new know-how can be obtained from those that have been examined, the viewpoint of data mining can be changed, new rules can be generated, and this information will be effective information for subsequent examination support.

It is a figure quoted for explaining the model which implements the medical fee statement information reading method and the examination support method in this embodiment, and the relationship with the use subject. FIG. 2 is a block diagram showing the internal configuration of a receipt management system 4 in which the medical remuneration statement information reading device according to the present embodiment is implemented by developing functions. FIG. 2 is a block diagram showing the internal configuration of a receipt management system 4 in which the medical fee statement examination support device according to the present embodiment is implemented by developing functions. It is a flowchart quoted for explaining the operation of the medical remuneration statement information reading device in the present embodiment. It is a flowchart quoted for explaining the operation of the medical fee statement examination support device in the present embodiment. It is a flowchart quoted for explaining the operation of the medical fee statement examination support device in the present embodiment. It is a figure quoted in order to explain the field which constitutes a medical remuneration statement used in this embodiment. It is a figure showing an example of a statement item of a medical remuneration statement used for correlation rule analysis used in this embodiment. It is a figure showing an example of the screen composition used by the medical treatment fee statement examination support device in this embodiment. It is a figure showing an example of the screen composition used by the medical treatment fee statement examination support device in this embodiment. It is a figure showing an example of the screen composition used by the medical treatment fee statement examination support device in this embodiment.

Explanation of reference numerals

4 Reception Management System 201 Image File 202 Character Reading Unit 203 Text File 204 Virtual Character Reading File 205 Reception Data Importing Unit 206 Reception Dictionary 207 Keyword File 208 Recording Medium File 209 Code Conversion Unit 210 Code Table 211 Code File 212 Reception Dictionary Generation Unit 213 Reception DB
214 Sorting rule selection unit 215 Recept sorting unit 216 Sorting rule DB
217 Judgment rule selection unit 218 Recept extraction unit 219 Judgment rule DB
220 Rule / Error Reception Image Display Unit 221 Sorting Condition Importing Unit 222 Sorting Rule Update Unit 223 Judgment Condition Importing Unit 224 Judgment Rule Updating Unit

Claims (16)

In examining the medical fee statement, a medical fee statement information reading device that takes in the item information described in the medical fee statement,
The medical remuneration statement is captured as an image, and the basic information column of the medical remuneration statement is read. For a character string described in an area that is difficult to recognize by ordinary reading, character line candidates are read and the character string is read. A character reading unit to cut out as
The combination of the parts of the character string cut out by the character reading unit and the frequently used character string described in the medical remuneration statement are compared with a medical remuneration statement dictionary described in advance, and the medical remuneration statement is compared. When there is a combination of character strings that match the dictionary, a medical remuneration statement keyword capturing unit that captures the character strings as keyword data;
A medical remuneration statement information reader characterized by comprising: The character reading unit includes:
After extracting a character line candidate and a character cutout candidate from the captured image, further storing the character cutout candidate as a virtual character reading file, storing a combination of character strings matching the medical remuneration statement dictionary as a file. The medical remuneration statement information reader according to claim 1, wherein:   3. A code conversion unit which codes the text data, medical fee statement data requested by a recording medium, and text data of basic information, and stores them in a common database. The medical remuneration statement information reader according to 1. When performing the examination of the items described in the medical fee statement, a medical fee statement examination support device that aims to improve the efficiency of the examination,
A sorting rule selection unit that displays sorting rules that have been input in the file in advance and selects a sorting rule to be applied from among them;
Applying the sorting rule selected in the sorting rule selection unit, a receipt sorting unit that sorts into a reception group having a large number of erroneous receipts and a reception group having a small number of erroneous claims,
A medical fee statement examination support device characterized by comprising: The sorting rule selection unit,
The medical fee statement examination support device according to claim 4, wherein a sorting rule to be applied is selected from sorting rules generated by data mining from past examination results or predefined sorting rules. .   The medical fee statement examination support device according to claim 4 or 5, further comprising a sorting rule updating unit that reflects the latest examination result in the sorting determination rule. When performing the examination of the items described in the medical fee statement, a medical fee statement examination support device that aims to improve the efficiency of the examination,
A keyword display unit that displays a keyword extracted from the items in the summary column and the name of the injury / illness column described in the medical remuneration statement by the character reading unit;
A determination rule selection unit having a function of prompting reading or input of a determination rule to be applied, and further having a function of selecting a read or input determination rule;
A medical remuneration statement display unit that displays an image of a medical remuneration statement that may contain an error by applying the selected determination rule to the extracted keyword,
A medical fee statement examination support device characterized by comprising: The medical fee statement display unit,
The medical fee statement examination support device according to claim 7, wherein a keyword that is a reason for being extracted as the medical fee statement that may include the error is highlighted. The medical fee statement display unit,
The medical fee statement examination support device according to claim 7, wherein the applied determination rule is displayed on a sticky note, and a message indicating a reason extracted when the sticky note is selected is displayed. The medical fee statement display unit,
10. A medical treatment fee statement that may include a corresponding error is extracted by designating a determination rule serving as an extraction condition, and is displayed as an image. Medical examination fee statement examination support device described in the section. In examining the medical fee statement, a medical fee statement information reading method that incorporates the item information described in the medical fee statement,
Capture the medical fee statement as an image, acquire text data in the basic information column, and read out character line candidates and character string cutouts as components for other items where character recognition is difficult.
Any combination of the cut-out of the read character strings and the frequently used character strings described in the medical remuneration statement are compared with the medical remuneration statement dictionary described in advance, and match with the medical remuneration statement dictionary. When there is a character string, import the character string as text data,
A medical fee statement information reading method characterized by the above-mentioned. A medical fee statement screening support method for improving the efficiency of the examination when performing the examination of the items described in the medical fee statement,
Using the contents of multiple items described in the medical fee reimbursement statement and the sorting rules obtained from past examination results, sort into a group of claims with many errors and a group of claims with few errors thing,
A medical fee statement review support method characterized by the following. A medical fee statement screening support method for improving the efficiency of the examination when performing the examination of the items described in the medical fee statement,
The abstract column described in the medical fee statement, the keyword extracted from the injury and disease name column is displayed,
A judgment rule set in advance is read or displayed on the screen by prompting the input of the judgment rule, a judgment rule to be applied is selected and determined, and the selected judgment rule is applied to the extracted keyword. Displaying medical reimbursement statements that may include errors by doing so,
A medical fee statement review support method characterized by the following. In examining the medical fee statement, a program used in a medical fee statement information reading device that takes in the item information described in the medical fee statement,
Capturing the medical fee statement as an image, and reading a character string in a basic information column of the medical fee statement;
For items for which character recognition is difficult, a step of cutting out a character string from the captured medical fee statement image and reading it as a part,
The parts of the cut-out character string are determined in all combinations, and the frequently used character strings described in the medical remuneration statement are compared with a previously described medical remuneration statement dictionary. Capturing the character string as text data when
The read text data, character string data read from the basic information column,
Encoding all the data of the medical reimbursement statement charged on magnetic media;
Medical fee statement information reading program that causes a computer to execute the procedure. When performing the examination of the items described in the medical remuneration statement, is a program used in the medical remuneration statement examination support device that aims to increase the efficiency of examination,
Displaying sorting rules created from past receipt data;
Prompting selection of a sorting rule to be applied from the displayed sorting rules,
By applying the selected / determined sorting rule, a step of sorting into a group of receipts with many erroneous claims and a group of less receipts with erroneous claims,
A medical fee statement examination support program that causes a computer to execute When performing the examination of the items described in the medical remuneration statement, is a program used in the medical remuneration statement examination support device that aims to increase the efficiency of examination,
Displaying the keyword extracted by the medical fee statement information reader;
Generated from the correlation between the keywords, prompting the reading or input of the determination rule, prompting the selection of the determination rule to be applied from the displayed rules, and displaying the selected determination rule,
Image displaying a medical remuneration statement that may include an error by applying the selected determination rule to the extracted keyword,
A medical fee statement examination support program that causes a computer to execute

Images