2019/12/16 - AWS Comprehend Medical - 2 new api methods
Changes New Ontology linking APIs will provides medication concepts normalization and Diagnoses codes from input text. In this release we will provide two APIs - RxNorm and ICD10-CM.
InferICD10CM detects medical conditions as entities listed in a patient record and links those entities to normalized concept identifiers in the ICD-10-CM knowledge base from the Centers for Disease Control.
See also: AWS API Documentation
Request Syntax
client.infer_icd10_cm(
Text='string'
)
string
[REQUIRED]
The input text used for analysis. The input for InferICD10CM is a string from 1 to 10000 characters.
dict
Response Syntax
{
'Entities': [
{
'Id': 123,
'Text': 'string',
'Category': 'MEDICAL_CONDITION',
'Type': 'DX_NAME',
'Score': ...,
'BeginOffset': 123,
'EndOffset': 123,
'Attributes': [
{
'Type': 'ACUITY'|'DIRECTION'|'SYSTEM_ORGAN_SITE'|'QUALITY'|'QUANTITY',
'Score': ...,
'RelationshipScore': ...,
'Id': 123,
'BeginOffset': 123,
'EndOffset': 123,
'Text': 'string',
'Traits': [
{
'Name': 'NEGATION'|'DIAGNOSIS'|'SIGN'|'SYMPTOM',
'Score': ...
},
]
},
],
'Traits': [
{
'Name': 'NEGATION'|'DIAGNOSIS'|'SIGN'|'SYMPTOM',
'Score': ...
},
],
'ICD10CMConcepts': [
{
'Description': 'string',
'Code': 'string',
'Score': ...
},
]
},
],
'PaginationToken': 'string',
'ModelVersion': 'string'
}
Response Structure
(dict) --
Entities (list) --
The medical conditions detected in the text linked to ICD-10-CM concepts. If the action is successful, the service sends back an HTTP 200 response, as well as the entities detected.
(dict) --
The collection of medical entities extracted from the input text and their associated information. For each entity, the response provides the entity text, the entity category, where the entity text begins and ends, and the level of confidence that Amazon Comprehend Medical has in the detection and analysis. Attributes and traits of the entity are also returned.
Id (integer) --
The numeric identifier for the entity. This is a monotonically increasing id unique within this response rather than a global unique identifier.
Text (string) --
The segment of input text that is matched to the detected entity.
Category (string) --
The category of the entity. InferICD10CM detects entities in the MEDICAL_CONDITION category.
Type (string) --
Describes the specific type of entity with category of entities. InferICD10CM detects entities of the type DX_NAME .
Score (float) --
The level of confidence that Amazon Comprehend Medical has in the accuracy of the detection.
BeginOffset (integer) --
The 0-based character offset in the input text that shows where the entity begins. The offset returns the UTF-8 code point in the string.
EndOffset (integer) --
The 0-based character offset in the input text that shows where the entity ends. The offset returns the UTF-8 code point in the string.
Attributes (list) --
The detected attributes that relate to the entity. An extracted segment of the text that is an attribute of an entity, or otherwise related to an entity, such as the nature of a medical condition.
(dict) --
The detected attributes that relate to an entity. This includes an extracted segment of the text that is an attribute of an entity, or otherwise related to an entity. InferICD10CM detects the following attributes: Direction , System, Organ or Site , and Acuity .
Type (string) --
The type of attribute. InferICD10CM detects entities of the type DX_NAME .
Score (float) --
The level of confidence that Amazon Comprehend Medical has that the segment of text is correctly recognized as an attribute.
RelationshipScore (float) --
The level of confidence that Amazon Comprehend Medical has that this attribute is correctly related to this entity.
Id (integer) --
The numeric identifier for this attribute. This is a monotonically increasing id unique within this response rather than a global unique identifier.
BeginOffset (integer) --
The 0-based character offset in the input text that shows where the attribute begins. The offset returns the UTF-8 code point in the string.
EndOffset (integer) --
The 0-based character offset in the input text that shows where the attribute ends. The offset returns the UTF-8 code point in the string.
Text (string) --
The segment of input text which contains the detected attribute.
Traits (list) --
The contextual information for the attribute. The traits recognized by InferICD10CM are DIAGNOSIS , SIGN , SYMPTOM , and NEGATION .
(dict) --
Contextual information for the entity. The traits recognized by InferICD10CM are DIAGNOSIS , SIGN , SYMPTOM , and NEGATION .
Name (string) --
Provides a name or contextual description about the trait.
Score (float) --
The level of confidence that Amazon Comprehend Medical has that the segment of text is correctly recognized as a trait.
Traits (list) --
Provides Contextual information for the entity. The traits recognized by InferICD10CM are DIAGNOSIS , SIGN , SYMPTOM , and NEGATION.
(dict) --
Contextual information for the entity. The traits recognized by InferICD10CM are DIAGNOSIS , SIGN , SYMPTOM , and NEGATION .
Name (string) --
Provides a name or contextual description about the trait.
Score (float) --
The level of confidence that Amazon Comprehend Medical has that the segment of text is correctly recognized as a trait.
ICD10CMConcepts (list) --
The ICD-10-CM concepts that the entity could refer to, along with a score indicating the likelihood of the match.
(dict) --
The ICD-10-CM concepts that the entity could refer to, along with a score indicating the likelihood of the match.
Description (string) --
The long description of the ICD-10-CM code in the ontology.
Code (string) --
The ICD-10-CM code that identifies the concept found in the knowledge base from the Centers for Disease Control.
Score (float) --
The level of confidence that Amazon Comprehend Medical has that the entity is accurately linked to an ICD-10-CM concept.
PaginationToken (string) --
If the result of the previous request to InferICD10CM was truncated, include the PaginationToken to fetch the next page of medical condition entities.
ModelVersion (string) --
The version of the model used to analyze the documents, in the format n .*n* .*n* You can use this information to track the model used for a particular batch of documents.
InferRxNorm detects medications as entities listed in a patient record and links to the normalized concept identifiers in the RxNorm database from the National Library of Medicine.
See also: AWS API Documentation
Request Syntax
client.infer_rx_norm(
Text='string'
)
string
[REQUIRED]
The input text used for analysis. The input for InferRxNorm is a string from 1 to 10000 characters.
dict
Response Syntax
{
'Entities': [
{
'Id': 123,
'Text': 'string',
'Category': 'MEDICATION',
'Type': 'BRAND_NAME'|'GENERIC_NAME',
'Score': ...,
'BeginOffset': 123,
'EndOffset': 123,
'Attributes': [
{
'Type': 'DOSAGE'|'DURATION'|'FORM'|'FREQUENCY'|'RATE'|'ROUTE_OR_MODE'|'STRENGTH',
'Score': ...,
'RelationshipScore': ...,
'Id': 123,
'BeginOffset': 123,
'EndOffset': 123,
'Text': 'string',
'Traits': [
{
'Name': 'NEGATION',
'Score': ...
},
]
},
],
'Traits': [
{
'Name': 'NEGATION',
'Score': ...
},
],
'RxNormConcepts': [
{
'Description': 'string',
'Code': 'string',
'Score': ...
},
]
},
],
'PaginationToken': 'string',
'ModelVersion': 'string'
}
Response Structure
(dict) --
Entities (list) --
The medication entities detected in the text linked to RxNorm concepts. If the action is successful, the service sends back an HTTP 200 response, as well as the entities detected.
(dict) --
The collection of medical entities extracted from the input text and their associated information. For each entity, the response provides the entity text, the entity category, where the entity text begins and ends, and the level of confidence that Amazon Comprehend Medical has in the detection and analysis. Attributes and traits of the entity are also returned.
Id (integer) --
The numeric identifier for the entity. This is a monotonically increasing id unique within this response rather than a global unique identifier.
Text (string) --
The segment of input text extracted from which the entity was detected.
Category (string) --
The category of the entity. The recognized categories are GENERIC or BRAND_NAME .
Type (string) --
Describes the specific type of entity. For InferRxNorm, the recognized entity type is MEDICATION .
Score (float) --
The level of confidence that Amazon Comprehend Medical has in the accuracy of the detected entity.
BeginOffset (integer) --
The 0-based character offset in the input text that shows where the entity begins. The offset returns the UTF-8 code point in the string.
EndOffset (integer) --
The 0-based character offset in the input text that shows where the entity ends. The offset returns the UTF-8 code point in the string.
Attributes (list) --
The extracted attributes that relate to the entity. The attributes recognized by InferRxNorm are DOSAGE , DURATION , FORM , FREQUENCY , RATE , ROUTE_OR_MODE , and STRENGTH .
(dict) --
The extracted attributes that relate to this entity. The attributes recognized by InferRxNorm are DOSAGE , DURATION , FORM , FREQUENCY , RATE , ROUTE_OR_MODE .
Type (string) --
The type of attribute. The types of attributes recognized by InferRxNorm are BRAND_NAME and GENERIC_NAME .
Score (float) --
The level of confidence that Comprehend Medical has that the segment of text is correctly recognized as an attribute.
RelationshipScore (float) --
The level of confidence that Amazon Comprehend Medical has that the attribute is accurately linked to an entity.
Id (integer) --
The numeric identifier for this attribute. This is a monotonically increasing id unique within this response rather than a global unique identifier.
BeginOffset (integer) --
The 0-based character offset in the input text that shows where the attribute begins. The offset returns the UTF-8 code point in the string.
EndOffset (integer) --
The 0-based character offset in the input text that shows where the attribute ends. The offset returns the UTF-8 code point in the string.
Text (string) --
The segment of input text which corresponds to the detected attribute.
Traits (list) --
Contextual information for the attribute. InferRxNorm recognizes the trait NEGATION for attributes, i.e. that the patient is not taking a specific dose or form of a medication.
(dict) --
The contextual information for the entity. InferRxNorm recognizes the trait NEGATION , which is any indication that the patient is not taking a medication.
Name (string) --
Provides a name or contextual description about the trait.
Score (float) --
The level of confidence that Amazon Comprehend Medical has in the accuracy of the detected trait.
Traits (list) --
Contextual information for the entity.
(dict) --
The contextual information for the entity. InferRxNorm recognizes the trait NEGATION , which is any indication that the patient is not taking a medication.
Name (string) --
Provides a name or contextual description about the trait.
Score (float) --
The level of confidence that Amazon Comprehend Medical has in the accuracy of the detected trait.
RxNormConcepts (list) --
The RxNorm concepts that the entity could refer to, along with a score indicating the likelihood of the match.
(dict) --
The RxNorm concept that the entity could refer to, along with a score indicating the likelihood of the match.
Description (string) --
The description of the RxNorm concept.
Code (string) --
RxNorm concept ID, also known as the RxCUI.
Score (float) --
The level of confidence that Amazon Comprehend Medical has that the entity is accurately linked to the reported RxNorm concept.
PaginationToken (string) --
If the result of the previous request to InferRxNorm was truncated, include the PaginationToken to fetch the next page of medication entities.
ModelVersion (string) --
The version of the model used to analyze the documents, in the format n .*n* .*n* You can use this information to track the model used for a particular batch of documents.