KR20220080364A - Method for providing service integrating medical report and FHIR mapping using deep learning - Google Patents

Abstract

The present invention relates to standardization of medical information and a method for integrating medical information through the standardization. A method for generating an FHIR resource based on heterogeneous medical data according to an embodiment of the present invention includes extracting field names and field values of data in collected heterogeneous medical information; determining a degree of similarity of the extracted field name and defining a mapping rule between the field name and a Fast Healthcare Interoperability Resources (FHIR) standard item according to the degree of similarity; and storing the field values according to the defined mapping rule, wherein when a mapping rule between the field name and a standard item is not defined, the similarity of the field values is determined, and the field name is set as standard. It is desirable to map it to an item. According to the present invention, it is possible to standardize various types of medical information through mapping of medical information conforming to the FHIR standard and provide big data-based medical services through this.

Description

FHIR medical information standard mapping using deep learning and a method for providing service integrating medical report and FHIR mapping using deep learning

The present invention relates to standardization of medical information and a method for integrating medical information through the standardization.

The difference in the way of expressing medical information managed according to various terminologies has become a barrier to the realization of current big data-based medical services, and accordingly, the need to integrate and manage medical information has been felt.

HL7 developed various standards such as V2 Message, V3 RIM, and CDA. Even if you are not a professional system developer, the need to understand data structures and implement services is increasing.

Therefore, HL7 developed the FHIR (Fast Healthcare Interoperability Resources) framework, which is freely extensible by users, while integrating standards to compensate for existing shortcomings and respond flexibly to changes.

However, in the actual medical market, additional work is required to map medical information directly to FHIR, and accordingly, various APIs are being additionally developed.

Using Azure API for FHIR, Microsoft is providing a service that enables rapid data exchange through the FHIR API supported by the cloud's managed Platform-as-a-Service (PaaS) offering.

Therefore, it is required to develop a mapping solution that can be dynamically applied based on an efficient mapping method and an actual use environment by applying deep learning in line with this flow.

An object of the present invention is to propose a method of mapping medical information that meets the FHIR standard.

An object of the present invention is to propose a mapping method without loss of medical information.

Another object of the present invention is to propose a method for standardizing by generating a rule capable of mapping undefined data using deep learning.

In accordance with an embodiment of the present invention for solving the above technical problem, a method for generating a FHIR resource based on heterogeneous medical data includes extracting field names and field values of data in collected heterogeneous medical information; determining a degree of similarity of the extracted field name and defining a mapping rule between the field name and a Fast Healthcare Interoperability Resources (FHIR) standard item according to the degree of similarity; and storing the field values according to the defined mapping rule, wherein when a mapping rule between the field name and a standard item is not defined, the similarity of the field values is determined, and the field name is set as standard. It is desirable to map it to an item.

The defining of the mapping rule may include: mapping the medical information to a first hierarchical item divided into a pathological examination item, a term system item, and a diagnosis code item; and mapping the first layer item to an FHIR standard item.

The pathological examination items are preferably generated with reference to LOINC (Logical Observation Identifiers Names and Codes).

The term system item is preferably created with reference to SNOMED CT (Systematized Nomenclature of Medicine Clinical Terms).

The diagnostic code item is preferably generated with reference to ICD (International classification of diseases)-10.

The method further includes storing the defined mapping rule, wherein the storing of the mapping rule includes receiving a user's confirmation on the new mapping rule for the field name according to the determination of the similarity of the field value, and setting the confirmed new mapping rule. It is desirable to reflect it in the defined mapping rule.

It is preferable that the similarity of the field name or field value is calculated through a neural network that performs semantic analysis of text, and the neural network is further trained using the similarity of the field value according to the confirmed new mapping rule.

The medical information includes a Personal Health Record (PHR),

Preferably, the personal health record is normalized based on an information acquisition period.

The personal health record is divided into a first personal health record based on EMR and a second personal health record according to the patient's individual self-diagnosis, and it is preferable that the second personal health record be given reliability according to the first personal health record do.

According to the present invention, it is possible to standardize various types of medical information through mapping of medical information conforming to the FHIR standard and provide big data-based medical services through this.

The present invention uses deep learning to generate a mapping rule according to the FHIR standard even for data for which a mapping rule is not defined, and through this, it is possible to flexibly cope with information collected from a new device or sensor.

In addition, the present invention can provide a management frame without data loss by including other information in the expandable area of the FHIR.

1 is a diagram illustrating an implementation example of a system in which a method for generating an FHIR resource based on heterogeneous medical data according to an embodiment of the present invention is operated.
2 is a diagram illustrating an FHIR structure according to an embodiment of the present invention.
3A and 3B are diagrams illustrating a flow of a system in which a method for generating an FHIR resource based on heterogeneous medical data according to an embodiment of the present invention is operated.
4 to 7 are diagrams illustrating an FHIR structure according to an embodiment of the present invention.
8 is a diagram illustrating a structure of a neural network in which a method for generating an FHIR resource based on heterogeneous medical data according to an embodiment of the present invention is operated.
9 is a diagram illustrating an intermediate first layer item for FHIR mapping according to an embodiment of the present invention.
10 is a diagram showing the configuration of a mapping server according to an embodiment of the present invention.

The following is merely illustrative of the principles of the invention. Therefore, those skilled in the art can devise various devices that, although not explicitly described or shown herein, embody the principles of the invention and are included in the spirit and scope of the invention. In addition, it should be understood that all conditional terms and examples listed herein are, in principle, expressly intended only for the purpose of understanding the inventive concept, and are not limited to the specifically enumerated embodiments and states. .

The above objects, features and advantages will become more apparent through the following detailed description in relation to the accompanying drawings, and accordingly, those of ordinary skill in the art to which the invention pertains will be able to easily practice the technical idea of the invention. .

In addition, in the description of the invention, if it is determined that a detailed description of a known technology related to the invention may unnecessarily obscure the gist of the invention, the detailed description thereof will be omitted. Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram illustrating a system for providing a method for generating an FHIR resource based on heterogeneous medical data according to an embodiment of the present invention.

The system according to this embodiment includes a standard mapping server 100 that collects various medical information 50 provided from a medical institution or a wearable device, and an information request server that receives information processed according to the target standard FHIR ( 200) consists of

That is, the standard mapping server 100 collects medical information from various institutions or devices and maps it according to the standard.

In addition to hospitals, institutions may include organizations that manage health insurance-related corporations or other disease information, and thus can provide coded information according to existing standards and provide additional judgment information as a separate item.

In addition, recently, devices incorporating various biometric sensors are being released in accordance with the increase in the Internet of Things and mobile communication markets and the miniaturization of sensors, through which personal health status information can be collected.

The sensed information may be the sensed value itself or diagnostic information diagnosed by the wearable device based on the sensed value.

However, since such a personalized device is not a judgment based on specialized medical knowledge, a separate data purification process may be performed, and for example, the judgment result of the highest/lowest blood pressure may be readjusted based on the value diagnosed at the hospital.

The standard mapping server 100 maps the medical information collected from the above various sources according to the FHIR.

FHIR is a standard that describes data formats and elements and application programming interfaces for exchanging electronic health records.

Specifically, referring to FIG. 2 , the standard targeted by the mapping server 100 is FHIR, which is divided into 5 items in Release 4, and each has a defined structure within the items. It is divided into 'Clinical, Diagnostics, and Medication' that can be administered, and within 'Medication', which includes drug-related content, immunity-related content is defined as 'Immunization'.

In other words, FHIR systematizes patient's medical data so that companies and individuals that use medical information in addition to various medical institutions can conveniently acquire health information and receive support for services such as telemedicine and prescriptions accordingly.

Accordingly, the standard mapping server 100 according to the present embodiment generates a mapping rule based on semantic interpretation using a stored rule or a neural network when storing the obtained medical information in a resource according to FHIR.

Hereinafter, a specific operation will be described with reference to FIG. 3A.

Referring to FIG. 3A , the standard mapping server 100 collects various standard or atypical medical information, and extracts a field name that defines data of the medical information and a field value that is an actual data value ( S100 ).

Accordingly, the mapping rule may be determined preferentially by referring to the field name, but if there is no field name rule, the mapping rule may be determined by sequentially referring to the field values.

That is, if there is a mapping rule predetermined according to a standard using the extracted information, direct mapping is performed.

Specifically, in the present embodiment, the mapping may be performed by mapping medical information corresponding to raw data to FHIR directly, instead of mapping it to an intermediate item, and then storing it as an FHIR standard.

However, when the mapping rule is not determined, the degree of similarity of the extracted field name is determined, and a mapping rule between the field name and the FHIR standard item is defined according to the degree of similarity (S200).

Referring to FIG. 3B , in order to define a mapping rule, medical information may be mapped to a first layer item, and FHIR resources may be made based on the mapped first layer item ( S210 ).

If there is no mapping rule, the standard mapping server 100 according to the present embodiment may directly define the rule based on semantic interpretation.

In this case, the newly defined mapping rule may be generated by referring to a data field name and a field value of the collected medical information, respectively.

That is, in this embodiment, a new mapping rule is redefined based on the semantic analysis of field names and field values through a neural network.

Specifically, when a similar field name exists in the existing mapping rule by using the degree of similarity between interpreted semantic features, the existing rule may be applied as it is.

On the other hand, if the similarity of the field name is less than or equal to the reference threshold and there is no applicable conventional rule, the field value is referred to.

That is, a mapping rule can be defined when similarity greater than or equal to the standard is calculated by determining the similarity with the standard item through the field name. S300).

In this case, based on the semantic interpretation of the field value, it is possible to find an item with a similar field value in the existing rule and refer to the rule that maps the upper field name in reverse.

In this case, the defined rule may be mapped to the first layer item instead of being set to directly correspond to the FHIR resource.

When mapped to the first hierarchical item, since the first hierarchical item is directly defined in the FHIR resource, the medical information input according to the FHIR is mapped ( S220 ).

The new mapping rule generated through the above process may be stored and used for mapping of medical information to be input later.

Also, although not additionally shown, the storing of the mapping rule includes receiving a user's confirmation on the new mapping rule for the field name according to the determination of the similarity of the field value, and reflecting the confirmed new mapping rule in the defined mapping rule. It is also possible to perform verification on

Hereinafter, the structure of the first hierarchical item according to the present embodiment will be described.

To this end, the FHIR structure is first described with reference to FIG. 4 . For example, in the case of a medication item, data is defined in the structure shown in FIG. 4 .

In this case, the code information is defined to be mapped using SNOMED code information, which is another standard. SNOMED is a standard for medical terminology that provides codes, terms, synonyms, and definitions that define clinical documentation and reporting systems. Therefore, a rule can be set so that the first layer item is preferentially mapped to a code according to SNOMED, and the mapped information can be referred to in 'Medication' according to the FHIR standard.

Also, referring to FIG. 5 , in the case of a Diagnostic Report item, data is defined in the same structure as in FIG. 5 .

In this case, the code information is defined to be mapped using LOINC code information, which is another standard. LOINC is a code defined to identify laboratory and clinical laboratory results.

Therefore, it allows the mapping server 100 according to the present embodiment to define the mapping rule as LOINC.

Specifically, for experimental results or numerical medical information, a rule mapping to LOINC can be created by referring to field names and field values, which can be stored and referenced in the diagnosis report in FHIR.

Referring to FIG. 6 , in addition to that, reference and usage rules of other standards are predefined in FHIR.

However, in the case of standard or unstructured data not defined here, it is necessary to create a separate mapping rule.

Referring to Figure 7, for example, also in relation to the family history (Familymemberhistory), the reasonCode is defined as using the above-described SNOMED, but died as a result (outcome) in the Condition according to the family history (deceased), permanent disability (permanent Disability) There are no rules defined for classification or death contribution (Contributedtodeath) field.

Therefore, the mapping server 100 can map to SNOMED based on the data definition criteria in the FHIR or the similarity with the past mapped data, but by recognizing the medical information related to the value related to the result of the family history or the death contribution, the neural network It is also possible to create a new mapping rule using

Specifically, provided medical information has metadata and actual information values as data about the data. The meaning of the field values can be inferred using a pre-trained neural network as the data field names and information values corresponding to the metadata. At this time, the meaning is output as each feature value using a neural network, and the corresponding feature value can be mapped using the similarity with the reference value learned for data definition.

Medical information whose rules have been set through the above process is directly mapped and stored in FHIR, but for other medical information, a mapping rule for the first hierarchical item is defined according to the similarity of field names and field values and stored as FHIR resources using the mapping rules. do (S400).

Referring to FIG. 8 , the neural network extracts feature information using a data field name and field value as input medical information. Layers inside the neural network are calculated to emphasize expressions or features with meaning in medical information. For example, when a CNN-based structure is created, features are further emphasized through convolution operation.

That is, the internal hidden layer may generate a first feature value for the data field name and a second feature value for the field value as intermediate feature information. The above feature values may be flattened and output as a fully connected layer of the output layer, and the corresponding output values may be classified based on the similarity with the items according to the first classification. Preferably, the first classification is divided into values according to the first hierarchical item, and the neural network may generate the first classification through the output value.

Referring to FIG. 9 , in the present embodiment, the first classification is divided into a pathological examination item, a term system item, and a diagnosis code item.

Specifically, the pathological examination item may be generated with reference to LOINC (Logical Observation Identifiers Names and Codes) in the above-described example.

The term system item may be generated with reference to SNOMED CT (Systematized Nomenclature of Medicine Clinical Terms).

The diagnostic code item may be generated by referring to International classification of diseases (ICD)-10. The above first hierarchical item may have a reference structure according to the FHIR therein.

LOINC is a naming system for tests and observations, for example, hemoglobin level, serum K (potassium), vital signs, intake/output, etc. Define physical observations.

Specific test items may be general chemical tests, blood tests, serological tests, microbiological tests, toxicity tests, antibiotic sensitivity tests, etc., and the results of the test items can be managed according to the items.

The items of SNOMED CT are defined in the table below, and the neural network learns the contents corresponding to the relation table. This creates a rule that maps medical information (see Table 1 below).

table name Explanation concept table
(Concepts Table) Each column in the concept table represents a clinical concept.
Each concept has a unique identifier for the concept and a Fully Specified Name that characterizes the concept's nature. synonym table
(Descriptions Table) The terms in the description table are related to the clinical concept and are used to represent the clinical concept.
Each term has an identifier, and includes the identifier of the concept the term expresses. relationship table
(Relationships Table) It shows the relationship between two concepts.
Relationship types are determined by the way in which other concepts are referenced.

ICD-10 is for disease classification based on causes rather than signs of disease, and classifies health status (V), cause of death (E), and tumor type (M). In addition, the International Classification of Diseases (ICD-) recommended by WHO 10) and Korea's KCD-7, which reflects the recent changes in the International Classification of Diseases oncology (ICD-O), are also referenced in parallel to reflect the medical system of each country and enable mapping.

In addition, if the medical information for which the rule is not determined as the first classification is reclassified and the frequency of medical information determined as the same classification is greater than or equal to a threshold, the neural network can create an additional item in the first classification, and through this, actively new data to be able to react to

That is, in the present embodiment, it is possible to respond to newly collected medical information by adding or modifying the first hierarchical item, which is an intermediate item, and to update the system in response to the scalability of the FHIR.

In addition, the rules inside the FHIR of the new item can be directly defined by the user, and medical information can be stored according to the FHIR through user confirmation.

Furthermore, in the present embodiment, medical information includes a personal health record (PHR) collected through a wearable device, and such personal health record may be normalized based on an information acquisition period.

That is, since the personal health record may have values in various units according to the type or version of the sensor used in the wearable device, it can be normalized and mapped.

In addition, it is possible to assign reliability to the personal health record according to whether it is verified by a specialized medical institution, and to map it by reflecting this.

That is, by dividing the personal health record into the EMR-based first personal health record and the second personal health record according to the patient's individual self-diagnosis, the second personal health record may be given reliability according to the first personal health record.

In the above example, when the blood pressure measured by the wearable smart band is used as it is, the second personal health record may be used, and information readjusted by comparing and analyzing the pulse waveform with the reference blood pressure may be managed as the first personal health record.

Referring to FIG. 10 , the above mapping server 100 may be configured as shown in FIG. 10 .

First, the medical information collection unit 110 collects various types of medical information 50 provided from the aforementioned medical institution or wearable device.

The medical information feature extractor 120 may define a new mapping rule by referring to a field name or field value of medical information in which a mapping rule is not defined using the neural network 125 .

Specifically, in this embodiment, the neural network 125 does not map directly according to the FIHR terminology, but first maps medical information to a classification according to the first hierarchical item, and maps the classified information to FHIR to minimize the loss of medical information. do.

The mapping rules generated through the above process are stored in the mapping rule storage unit 130, and the mapping rule storage unit also stores the rules according to the FHIR terminology 135 in advance to determine whether there is a mapping rule for the input medical information. to avoid duplicate definitions.

The data providing unit 140 generates and outputs an FHIR resource. Therefore, the information request server 200 that receives the information processes and processes the collected data based on FHIR so that it can be utilized in various fields such as diagnosis and insurance calculation.

As described above, according to the present invention, it is possible to standardize various types of medical information through mapping of medical information conforming to the FHIR standard and provide big data-based medical services through this.

The present invention uses deep learning to generate a mapping rule according to the FHIR standard even for data for which a mapping rule is not defined, and through this, it is possible to flexibly cope with information collected from a new device or sensor.

In addition, the present invention can provide a management frame without data loss by including other information in the expandable area of the FHIR.

Furthermore, various embodiments described herein may be implemented in a computer-readable recording medium using, for example, software, hardware, or a combination thereof.

According to the hardware implementation, the embodiments described herein include ASICs (application specific integrated circuits), DSPs (digital signal processors), DSPDs (digital signal processing devices), PLDs (programmable logic devices), FPGAs (field programmable gate arrays, It may be implemented using at least one of processors, controllers, micro-controllers, microprocessors, and other electrical units for performing functions. The described embodiments may be implemented in the control module itself.

According to the software implementation, embodiments such as the procedures and functions described in this specification may be implemented as separate software modules. Each of the software modules may perform one or more functions and operations described herein. The software code may be implemented as a software application written in a suitable programming language. The software code may be stored in the memory module and executed by the control module.

The above description is merely illustrative of the technical idea of the present invention, and those of ordinary skill in the art to which the present invention pertains may make various modifications, changes and substitutions within the scope without departing from the essential characteristics of the present invention. will be.

Accordingly, the embodiments disclosed in the present invention and the accompanying drawings are intended to explain, not to limit the technical spirit of the present invention, and the scope of the technical spirit of the present invention is not limited by these embodiments and the accompanying drawings . The protection scope of the present invention should be construed by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

Claims (11)

extracting field names and field values of data in the collected heterogeneous medical information;
determining a degree of similarity of the extracted field name and defining a mapping rule between the field name and a Fast Healthcare Interoperability Resources (FHIR) standard item according to the degree of similarity; and
and storing the field values according to the defined mapping rule,
When the mapping rule between the field name and the standard item is not defined, the field name is mapped to the standard item by determining the similarity of the field value. The method of claim 1,
The step of defining the mapping rule comprises:
mapping the medical information to a first hierarchical item divided into a pathological examination item, a term system item, and a diagnostic code item; and
and mapping the first layer item to an FHIR standard item. 3. The method of claim 2,
The pathological examination item is a heterogeneous medical data-based FHIR resource creation method, characterized in that it is generated with reference to LOINC (Logical Observation Identifiers Names and Codes). 3. The method of claim 2,
The term system item is a heterogeneous medical data-based FHIR resource generation method, characterized in that it is generated with reference to SNOMED CT (Systematized Nomenclature of Medicine Clinical Terms). 3. The method of claim 2,
The diagnostic code item is a FHIR resource generating method based on heterogeneous medical data, characterized in that it is generated with reference to ICD (International classification of diseases)-10. The method of claim 1,
Further comprising the step of storing the defined mapping rule,
The step of storing the mapping rule includes receiving a user's confirmation on the new mapping rule for the field name according to the determination of the similarity of the field value, and reflecting the confirmed new mapping rule in the defined mapping rule. Way. 7. The method of claim 6,
The similarity of the field name or field value is calculated through a neural network that performs semantic analysis of text,
The FHIR resource creation method, characterized in that the neural network is further trained using the similarity of field values according to the confirmed new mapping rule. The method of claim 1,
The medical information includes a Personal Health Record (PHR),
The personal health record is a heterogeneous medical data-based FHIR resource creation method, characterized in that normalized based on the information acquisition period. 9. The method of claim 8,
The personal health record is divided into a first personal health record based on EMR and a second personal health record according to the patient's individual self-diagnosis,
The method for generating a FHIR resource based on heterogeneous medical data, characterized in that the second personal health record is given reliability according to the first personal health record. a medical information collection unit that collects medical information;
Extracting field names and field values of data in the collected heterogeneous medical information, determining a similarity between a predetermined first hierarchical item and characteristic information of the extracted field name using a neural network, and determining the field name and FHIR according to the similarity (Fast Healthcare Interoperability Resources) A medical information feature extraction unit that defines a mapping rule between standard items;
a mapping rule storage unit for storing the defined mapping rule;
It characterized in that it comprises a data providing unit for providing the FHIR resource generated according to the mapping rule,
The medical information feature extraction unit determines the similarity of the field value when a mapping rule between a field name and a standard item is not defined, and maps the field name to a standard item. A program stored in a computer-readable recording medium for performing the method for generating a FHIR resource based on heterogeneous medical data according to claim 1 .

Images