Pipeline to Resolve Medication Codes(Transform)

Description

A pretrained resolver pipeline to extract medications and resolve their adverse reactions (ADE), RxNorm, UMLS, NDC, SNOMED CT codes, and action/treatments in clinical text.

Action/treatments are available for branded medication, and SNOMED codes are available for non-branded medication.

This pipeline can be used with Spark transform. You can use medication_resolver_pipeline as Lightpipeline (with annotate/fullAnnotate).

Copy S3 URI

How to use


from sparknlp.pretrained import PretrainedPipeline

medication_resolver_pipeline = PretrainedPipeline("medication_resolver_transform_pipeline", "en", "clinical/models")

text = """The patient was prescribed Amlodopine Vallarta 10-320mg, Eviplera. The other patient is given Lescol 40 MG and Everolimus 1.5 mg tablet."""

data = spark.createDataFrame([[text]]).toDF("text")

result = medication_resolver_pipeline.transform(data)


import com.johnsnowlabs.nlp.pretrained.PretrainedPipeline

val medication_resolver_pipeline = new PretrainedPipeline("medication_resolver_transform_pipeline",, "en", "clinical/models")

val data = Seq("""The patient was prescribed Amlodopine Vallarta 10-320mg, Eviplera. The other patient is given Lescol 40 MG and Everolimus 1.5 mg tablet.""").toDS.toDF("text")

val result = medication_resolver_pipeline.fit(data).transform(data)

Results

+----------------------------+---------+---------------------------+-------+--------------------------+------------------------------------------+--------+---------+-----------+-------------+
|chunk                       |ner_label|ADE                        |RxNorm |Action                    |Treatment                                 |UMLS    |SNOMED_CT|NDC_Product|NDC_Package  |
+----------------------------+---------+---------------------------+-------+--------------------------+------------------------------------------+--------+---------+-----------+-------------+
|Amlodopine Vallarta 10-320mg|DRUG     |Gynaecomastia              |722131 |NONE                      |NONE                                      |C1949334|425838008|00093-7693 |00093-7693-56|
|Eviplera                    |DRUG     |Anxiety                    |217010 |Inhibitory Bone Resorption|Osteoporosis                              |C0720318|NONE     |NONE       |NONE         |
|Lescol 40 MG                |DRUG     |NONE                       |103919 |Hypocholesterolemic       |Heterozygous Familial Hypercholesterolemia|C0353573|NONE     |00078-0234 |00078-0234-05|
|Everolimus 1.5 mg tablet    |DRUG     |Acute myocardial infarction|2056895|NONE                      |NONE                                      |C4723581|NONE     |00054-0604 |00054-0604-21|
+----------------------------+---------+---------------------------+-------+--------------------------+------------------------------------------+--------+---------+-----------+-------------+

Model Information

Model Name: medication_resolver_transform_pipeline
Type: pipeline
Compatibility: Healthcare NLP 5.2.1+
License: Licensed
Edition: Official
Language: en
Size: 3.4 GB

Included Models

  • DocumentAssembler
  • SentenceDetectorDLModel
  • TokenizerModel
  • WordEmbeddingsModel
  • MedicalNerModel
  • NerConverterInternalModel
  • TextMatcherInternalModel
  • ChunkMergeModel
  • ChunkMapperModel
  • ChunkMapperModel
  • ChunkMapperFilterer
  • Chunk2Doc
  • BertSentenceEmbeddings
  • SentenceEntityResolverModel
  • ResolverMerger
  • Doc2Chunk
  • ResolverMerger
  • ChunkMapperModel
  • ChunkMapperModel
  • ChunkMapperModel
  • Doc2Chunk
  • ChunkMapperModel
  • ChunkMapperModel
  • ChunkMapperModel
  • Finisher