JP2004318875A - Reading of medical record image - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To streamline data input and correction even outside a medical institution and ensure image reading allowing recombination by masking information for personal identification and imaging only a medical attribute information description part when a medical record is read with an image scanner. <P>SOLUTION: Digital data read from a medical record are masked in a specified data part for personal identification, and the remaining medical attribute information part is captured and saved as image data. For data that must be recombined, a personal identification code is recognized and encrypted, and the encrypted personal identification code is embedded in the masked region or a margin area of the medical record before saving or outputting. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention selectively digitizes biochemical test data, clinical data, and physical measurement data when digitizing personal identification information such as medical information including personal information as image data, and outputs the resulting image data. The present invention relates to an image reading technique in which an individual cannot be directly identified from an image.

  In data mining, which searches for the characteristics and trends of diseases from a large amount of patient data, data is input from a primitive form such as a medical record to an electronic medical record system using a keyboard or the like. In general, source forms include primary information that can identify an individual by combining only one or several pieces of information such as an individual's name, address, and telephone number (hereinafter referred to as first information), inspection data for each individual, Attribute information (hereinafter referred to as second information) necessary for secondary medical treatment such as clinical data is described.

  Because the medical record contains the first information that can easily identify an individual, when digitizing such data, work in a medical institution such as a hospital so that third parties do not see it. In terms of information security, the management of work places and data was very severely restricted. In addition, the head of the medical site supervised the work.

  Normally, when such primitive forms are digitized, a check box is displayed on the screen while using electronic medical record information management software (hereinafter referred to as electronic medical record) in a limited work environment in a medical institution such as a hospital. And input each item directly using the mouse and keyboard in the text box.

  In addition, the digitized data separates the first information and the second information that are easy to identify individuals, and can be reconnected to the first information if necessary, or can be connected again to the first information, or the first information again. Second information is output as non-reconnectable information that is output in a state where reconnection is not possible, and is used for statistical processing and data mining. In addition, when it was necessary to analyze the data as time-series data, each time a medical institution with primitive forms entered the data under a certain supervision while maintaining consistency with the previous data .

An electronic medical record system that takes personal information into consideration is disclosed in Japanese Patent Application Laid-Open No. 11-353404. This is a medical institution that stores the results of genetic testing of patients with the aim of providing an electronic medical record system that can search medical records of other patients with similar congenital constitutions for medical treatment support information from medical records. DB server is connected via a network, search processing of all genetic test results is performed from all clients, and the patient's medical record is divided into personal information files, medical record files, and genetic test result files for each item to protect personal information And performs a process of acquiring a medical record based on the position information of the medical record of the search result with only the searchable target as a group of genetic test result files. Searchable.
JP-A-11-353404

  However, in the related art, since data is input using a medical chart including first information, the data must be input at a medical site itself such as a hospital so as not to be seen by a third party. Therefore, there is a problem that it is difficult to secure a sufficient input space to perform the operation. In addition, since the first information related to personal identification is included, the work supervision must be performed by a person who is directly engaged in medical treatment, such as a physician at the medical site, and is not compared to the qualitative content of the work. Therefore, there was a problem that the time cost required for the supervisor was large.

  In addition, even in the work of confirming the digitized data, it is necessary to perform a calibration work by matching with the original form, so in order to prioritize medical care, to digitize the original form such as paper media, which is the original information, There has been a problem that a burden such as a drastic change in the flow of ordinary medical work has been imposed. In addition, since these medical data and medical examination data are data acquired in time series, they must be reconnected to data after a certain period of time, and each time the data is input by referring to the source form. Needed.

  Therefore, an object of the present invention is to provide a data processing technique which can request an external institution to perform the work while protecting personal information when digitizing the source form. is there.

  According to the present invention, in order to solve the above-described problems, a source form is read by an image scanner, and first information in which personal information is recorded is masked (including deletion of data) in accordance with a separately given instruction. (2) A method for saving or outputting only the image data of the information part, taking the data to an ethically managed joint research place outside the medical institution that performs statistical processing, and performing the work of digitizing the medical record information And an apparatus.

  Also, it recognizes the personal identification information described in the source form, further calculates an encrypted personal identification code by encrypting them, and stores a part of the image data of the second information or the area of the deleted first information. By superimposing and outputting an encrypted personal identification symbol on the data, a means for embedding reconnectable information as image data in the entire data is provided.

  According to the above-described means, it is possible to handle only the second information in which an individual cannot be specified by an external organization or the like in consideration of ethics, and to input data efficiently with a large number of human resources. In addition, since the image data on which the calibration work is based is at hand, the work can be performed without affecting the medical site. Further, since the personal identification symbol is encrypted and output together with the second information data, it is possible to proceed with digitization as data that can be reconnected. The storage medium according to the present invention includes various memory media such as a CD-ROM.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is an overall configuration diagram for explaining the present invention. The system shown in FIG. 1 includes a monitor, an information processing device (personal computer), various input devices (keyboard, scanner, etc.), and a communication device (modem, etc.). May be connected via the Internet. The data processing institution reads necessary items from image data (characters and images such as examination, diagnosis, treatment progress, and medication history) sent from the hospital and manually inputs the data into the electronic medical record system.

  In the present embodiment, digital conversion means 10 for converting a chart of a paper medium into image data and first information for inputting coordinate information and the like for recognizing a portion where the first information on the chart is described. Recognition designation data input means 11, data storage means 12 for storing chart designation data and recognition designation data for the first information recognition, first information masking means 13 for masking first information, image data of the chart It comprises a second information image data storage means 14 consisting of only the second information from which the first information has been removed, and a second information image data output means 15 for outputting the stored second information image data.

  Next, the operation of this embodiment will be described. First, a medical institution performs medical treatment using a medical chart as shown in FIG. As shown in the figure, information (hereinafter, referred to as first information) for specifying the individual such as the name, address, date of birth, and telephone number of the individual is written in the chart. In addition to the first information, attribute information necessary for analysis (hereinafter, referred to as second information) is written as medical data. Such a chart corresponds to a primitive form for digitization.

  This primitive form is bit-mapped using digital conversion means 10 such as an image scanner to obtain image data. Further, the image data is stored and stored by the data storage means 12. In some cases, the data may be compressed and restored, if necessary.

  Next, a coordinate position for masking the first information of the medical chart is input from the first information recognition designation data input 11. The first information masking process is performed at step 13 from the chart image data read from the data storage means, using the coordinate data in which the outline for the masking is designated by a polygonal line. FIG. 4 shows an example of the result of performing the masking process.

  Column data related to personal identification is painted in a uniform color, here white, as indicated by 40, 41, and 42 in FIG. Masking the personally identifiable data in this way makes it possible to create image data having second information that is difficult to reconnect to an individual.

  Further, the second information image data composed of only the second information obtained by masking the first information is stored in an auxiliary storage device such as a hard disk by randomizing the order of the image data in 14 second information image data storage means. It is stored on a medium (not shown). The second information image data output 15 allows the stored second information image data to be retrieved, retrieved, and displayed as needed.

  The data in the storage medium is sent to an external data entry company using a communication path, or is input to a portable storage medium and sent to an external organization. An external data input organization reads an identification code such as a barcode, and manually inputs necessary data to the electronic medical chart while viewing the medical chart image corresponding to the code.

  Next, a second embodiment will be described. FIG. 3 is a block diagram relating to the second embodiment. As in the first embodiment, the chart is digitized by the digital conversion means 10. Also, the masking coordinate position is input from the first information recognition designation data input unit 11. The image read by the digital conversion means 10 is stored in the data storage means 12.

  Here, in the second embodiment, the personal identification symbol recognition encrypting means 16 recognizes the personal identification data from the image data of the first information for the digitally converted image. In the present embodiment, for example, the personal identification data is obtained by directly printing the patient's code as a barcode or pasting a sticker on which the barcode is printed. The barcode position is recognized, and the barcode is recognized. As the barcode, a one-dimensional barcode used in physical distribution such as Code39 or Code128 or a two-dimensional barcode in which dots are arranged in a matrix are convenient.

  After recognizing the barcode, the personal identification symbol is encrypted. As the encryption processing, for example, RSA public key encryption (IEEE 1363) can be used. In addition, there is the MD5 system. In the case of linkable anonymization, patient code is encrypted with this code and name. The MD5 method is one of the hash functions (one-way summarization function) used for authentication and digital signatures. A fixed-length hash value is generated based on the original text and compared at both ends of the communication path. It is possible to detect whether the original text has been falsified on the way. Since the calculation method includes an irreversible one-way function with initial value sensitivity, the hash value takes a value like a pseudo-random number, and the original text cannot be reproduced based on the value. It is also very difficult to create another message that generates the same hash value.

  Further, the first information masking and encrypted personal identification information providing means 17 masks the first information as in the first embodiment. In addition, an encrypted personal identification symbol is embedded in the masked area or blank area. For the embedding method, in addition to embedding by overwriting the image data by converting it to a character string or barcode, refer to the image data file name and the encrypted personal identification symbol as one-to-one correspondence list data for later digitization It is also possible to write to another file if possible.

  The image data to which the encrypted personal identification information has been added is stored and stored in the storage unit 14 as in the first embodiment. Further, it can be displayed on the second information image data output 15 or printed on a paper medium.

  In this embodiment, the method of overwriting the encrypted personal identification information on the image data with a character string or a bar code has been described. An image watermarking method for restoring a personal identification code by a specific calculation method while no significant change is recognized can be used. Note that an IC tag can be used instead of the barcode. In the above-described embodiment, the present invention is applied to medical information. However, the present invention can also be applied to information relating to attributes of each individual such as financial information.

  As described above, according to the present embodiment, since only the second information is imagined and read while avoiding the primary information for identifying the individual of the medical record information, which is the source information, if the medical ethics is within an appropriate organization, it is distributed. Thus, data input can be performed efficiently by many people. In addition, since it is possible to keep the data as faithful to the original data at hand, even if the calibration work is performed after the data input, no extra burden is imposed on the medical site.

  In addition, by recognizing and encrypting the personal identification information, the image data itself can be managed in a reconnectable state. Therefore, even when the time-series data information is sequentially digitized, the connection with the update data is possible. It has an excellent feature that it is easy.

  As described above, according to the present invention, masking the first information on an individual, separating the information necessary for data mining such as test data and clinical data, and reading the image, the data which is difficult to reconnect is obtained. By acquiring or encrypting the personal identification code and outputting it together with the second information, it is possible to obtain image data of the second information that can be reconnected.

  By using these methods, in the medical field, the image data is read in a short time, and only the second information of the result is stored at a specific ethically controlled research site outside the medical institution using the stored data. Data entry can be performed while taking necessary time. In addition, since the image of the second information can be kept at a third-party external organization that is not a medical site, even if input errors are sufficiently confirmed and calibrated, there is no burden on the medical site. And provide a reliable and reliable data entry environment.

  In this way, only information related to the personal name is input at the medical institution, and then the anonymized information can be ordered to the outside, and the input operation can be requested. Therefore, the input operation is outsourced while protecting the personal information. be able to. Further, the encryption makes it difficult to allocate to the original data, and a random key is allocated to encrypt the patient's personal information, thereby enabling continuous code allocation.

FIG. 1 is an overall block diagram for explaining an embodiment of the present invention. FIG. 4 is a diagram for explaining an example of a medical chart to be input in the system according to the embodiment. FIG. 6 is a block diagram according to a second embodiment of the present invention. The figure for demonstrating an example of the chart image data which masked the 1st information.

Explanation of reference numerals

10 Digital conversion means 11 First information recognition designation data input means 12 Data storage means 13 First information masking means 14 Second information image data storage means 15 Second information image data output means 16 Personal identification symbol recognition encryption means 17 First information masking and encrypted personal identification information providing means

Claims (7)

  Digital conversion means for digitizing the source information in which the first information specifying the individual and the second information obtained as the medical attribute value of the individual exist, and digitally converted data A first information recognition designation data input means for separating and recognizing the first information from the source information; and a first information recognition designation from the conversion result data from the digital conversion means. First information masking means for masking first information according to the data, second information image data storage means for storing the second information image data after the first information masking, and externally storing the second information image data. And a second information image data output means for outputting to a medical record image.   The medical record image reading device according to claim 1, wherein the personal identification symbol encrypting means for converting a personal identification symbol described as the first information on the source information by a desired conversion method, A chart image reading apparatus, further comprising an encrypted personal identification information writing unit for embedding encrypted personal identification information obtained from the identification symbol encryption unit at a predetermined position of the second information image data.   A digital conversion process for digitizing the source information on a recording medium on which the source information having the first information specifying the individual and the second information obtained as the attribute value of the individual is recorded. A data storage step for storing digitally converted data; a first information recognition designating data input step for separating and recognizing first information from the source information; and a conversion result data from the digital conversion step. A first information masking step of masking the first information according to the first information recognition designation data, and a second information image data storing step of storing the second information image data after the first information masking; A second information image data output step for outputting the second information image data to the outside, and a recording medium in which personal attribute information is described. How to read the image.   A personal identification code encrypting step of converting the personal identification code described as the first information on the source information by a desired conversion method; and encrypted personal identification information obtained from the personal identification code encryption step. 4. The method according to claim 3, further comprising a step of writing encrypted personal identification information for embedding the information in a predetermined position of the second information image data.   A program for causing a computer to execute all steps or at least one step according to claim 3 or 4.   A storage medium storing the program according to claim 5.   In a data processing method for converting a form in which information identifying a person and attribute information of the person are recorded into an image file, a position where the personal identification information of the form is described is designated, and data in the designated range is masked. And data processing method.

Images