Pipeline to Resolve ICD-10-CM Codes

Description

This pipeline can extract clinical conditions, and map the clinical conditions to their respective ICD-10-CM codes using sbiobert_base_cased_mli Sentence Bert Embeddings. Users can refer to the following entity labels for pertinent concepts: ICD-10-CM entities: PROBLEM, CEREBROVASCULAR_DISEASE, COMMUNICABLE_DISEASE, DIABETES, DISEASE_SYNDROME_DISORDER, EKG_FINDINGS, HEART_DISEASE, HYPERLIPIDEMIA, HYPERTENSION, IMAGINGFINDINGS, INJURY_OR_POISONING, KIDNEY_DISEASE, OBESITY, ONCOLOGICAL, OVERWEIGHT, PREGNANCY, PSYCHOLOGICAL_CONDITION, SYMPTOM, VS_FINDING

Live Demo Open in Colab Download Copy S3 URI

How to use

from sparknlp.pretrained import PretrainedPipeline

resolver_pipeline = PretrainedPipeline("icd10cm_resolver_pipeline", "en", "clinical/models")

result = resolver_pipeline.fullAnnotate("""A 28-year-old female with a history of gestational diabetes mellitus diagnosed eight years and anisakiasis. Also, it was reported that fetal and neonatal hemorrhage.""")

import com.johnsnowlabs.nlp.pretrained.PretrainedPipeline

val result = resolver_pipeline.fullAnnotate("""A 28-year-old female with a history of gestational diabetes mellitus diagnosed eight years and anisakiasis. Also, it was reported that fetal and neonatal hemorrhage.""")

Results

|   |                        chunks | entities | icd10cm_code |
|--:|------------------------------:|---------:|-------------:|
| 0 | gestational diabetes mellitus |  PROBLEM |        O24.4 |
| 1 |                   anisakiasis |  PROBLEM |        B81.0 |
| 2 | fetal and neonatal hemorrhage |  PROBLEM |        P54.5 |

Model Information

Included Models

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