Drug Route Text Matcher Pipeline

Description

This pipeline, extracts drug administration route entities in clinical text. It recognizes routes including oral, IV, IM, subcutaneous, topical, transdermal, inhalation, sublingual, rectal, and more.

Copy S3 URI

How to use


from sparknlp.pretrained import PretrainedPipeline

pipeline = PretrainedPipeline("drug_route_matcher_pipeline", "en", "clinical/models")

sample_text = """ Metformin 500mg orally twice daily. Insulin glargine subcutaneously at bedtime. Morphine 4mg IV for pain. Albuterol via inhalation every 4 hours PRN."""

result = pipeline.transform(spark.createDataFrame([[sample_text]]).toDF("text"))


from johnsnowlabs import nlp, medical

pipeline = nlp.PretrainedPipeline("drug_route_matcher_pipeline", "en", "clinical/models")

sample_text = """ Metformin 500mg orally twice daily. Insulin glargine subcutaneously at bedtime. Morphine 4mg IV for pain. Albuterol via inhalation every 4 hours PRN."""

result = pipeline.transform(spark.createDataFrame([[sample_text]]).toDF("text"))


import com.johnsnowlabs.nlp.pretrained.PretrainedPipeline

val pipeline = PretrainedPipeline("drug_route_matcher_pipeline", "en", "clinical/models")

val sample_text = """ Metformin 500mg orally twice daily. Insulin glargine subcutaneously at bedtime. Morphine 4mg IV for pain. Albuterol via inhalation every 4 hours PRN."""

val result = pipeline.transform(spark.createDataFrame([[sample_text]]).toDF("text"))

Results


| chunk      | begin | end | label      |
| :--------- | ----: | --: | :--------- |
| orally     |    16 |  21 | DRUG_ROUTE |
| IV         |    93 |  94 | DRUG_ROUTE |
| inhalation |   120 | 129 | DRUG_ROUTE |

Model Information

Model Name: drug_route_matcher_pipeline
Type: pipeline
Compatibility: Healthcare NLP 6.3.0+
License: Licensed
Edition: Official
Language: en
Size: 937.8 KB

Included Models

  • DocumentAssembler
  • SentenceDetector
  • TokenizerModel
  • TextMatcherInternalModel
  • ChunkConverter