Extract Income and Social Status Entities from Social Determinants of Health Texts

Description

SDOH NER model is designed to detect and label social determinants of health (SDOH) income and social status-related entities within text data. Social determinants of health are crucial factors that influence individuals’ health outcomes, encompassing various social, economic, and environmental elements. The model has been trained using advanced machine-learning techniques on a diverse range of text sources. The model’s accuracy and precision have been carefully validated against expert-labeled data to ensure reliable and consistent results. Here are the labels of the SDOH NER model with their description:

Education:Patient’s educational background
Employment: Patient or provider occupational titles.
Financial_Status: Financial status refers to the state and condition of the person’s finances. “financial decline, debt, bankruptcy, etc.”
Income: Information regarding the patient’s income
Marital_Status: Terms that indicate the person’s marital status.
Population_Group: The population group that a person belongs to, that does not fall under any other entities. “refugee, prison patient, etc.”

Predicted Entities

Education, Employment, Financial_Status, Income, Marital_Status, Population_Group

Live Demo Open in Colab Download Copy S3 URI

How to use

from pyspark.sql.types import StringType

document_assembler = DocumentAssembler()\
    .setInputCol("text")\
    .setOutputCol("document")

sentence_detector = SentenceDetectorDLModel.pretrained("sentence_detector_dl", "en")\
    .setInputCols(["document"])\
    .setOutputCol("sentence")

tokenizer = Tokenizer()\
    .setInputCols(["sentence"])\
    .setOutputCol("token")

clinical_embeddings = WordEmbeddingsModel.pretrained("embeddings_clinical", "en", "clinical/models")\
    .setInputCols(["sentence", "token"])\
    .setOutputCol("embeddings")

ner_model = MedicalNerModel.pretrained("ner_sdoh_income_social_status", "en", "clinical/models")\
    .setInputCols(["sentence", "token", "embeddings"])\
    .setOutputCol("ner")

ner_converter = NerConverterInternal()\
    .setInputCols(["sentence", "token", "ner"])\
    .setOutputCol("ner_chunk")

pipeline = Pipeline(stages=[
    document_assembler, 
    sentence_detector,
    tokenizer,
    clinical_embeddings,
    ner_model,
    ner_converter   
    ])

sample_texts = ["Pt is described as divorced and pleasant when approached but keeps to himself. Pt is working as a plumber, but he gets financial diffuculties. He has a son student at college. He is an immigrant. He has a low income.", "Smith is a 55 years old, divorced Mexican American woman with financial problems.", "He lives with his family, in his own house in a remote town, with a monthly income of $1200 per month.","She reports feeling unsafe at home but is financially dependent.", "35-year-old, male patient, married, Asian immigrant, morbidly obese with a BMI of 32, and a history of hypertension presents to the clinic with worsening blood pressure readings. He is a high school graduate and is construction worker by profession."]

data = spark.createDataFrame(sample_texts, StringType()).toDF("text")

result = pipeline.fit(data).transform(data)

val document_assembler = new DocumentAssembler()
    .setInputCol("text")
    .setOutputCol("document")

val sentence_detector = SentenceDetectorDLModel.pretrained("sentence_detector_dl", "en")
    .setInputCols("document")
    .setOutputCol("sentence")

val tokenizer = new Tokenizer()
    .setInputCols("sentence")
    .setOutputCol("token")

val clinical_embeddings = WordEmbeddingsModel.pretrained("embeddings_clinical", "en", "clinical/models")
    .setInputCols(Array("sentence", "token"))
    .setOutputCol("embeddings")

val ner_model = MedicalNerModel.pretrained("ner_sdoh_income_social_status", "en", "clinical/models")
    .setInputCols(Array("sentence", "token", "embeddings"))
    .setOutputCol("ner")

val ner_converter = new NerConverterInternal()
    .setInputCols(Array("sentence", "token", "ner"))
    .setOutputCol("ner_chunk")

val pipeline = new Pipeline().setStages(Array(
    document_assembler, 
    sentence_detector,
    tokenizer,
    clinical_embeddings,
    ner_model,
    ner_converter   
))

val data = Seq(Array("Pt is described as divorced and pleasant when approached but keeps to himself. Pt is working as a plumber, but he gets financial diffuculties. He has a son student at college. He is an immigrant. He has a low income.","Smith is a 55 years old, divorced Mexican American woman with financial problems.", "He lives with his family, in his own house in a remote town, with a monthly income of $1200 per month.","She reports feeling unsafe at home but is financially dependent.", "35-year-old, male patient, married, Asian immigrant, morbidly obese with a BMI of 32, and a history of hypertension presents to the clinic with worsening blood pressure readings. He is a high school graduate and is construction worker by profession.")).toDS.toDF("text")

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

Results

+---------------------------------+-----+---+----------------+
|chunk                            |begin|end|ner_label       |
+---------------------------------+-----+---+----------------+
|divorced                         |19   |26 |Marital_Status  |
|plumber                          |98   |104|Employment      |
|financial diffuculties           |119  |140|Financial_Status|
|student                          |156  |162|Education       |
|college                          |167  |173|Education       |
|immigrant                        |185  |193|Population_Group|
|low income                       |205  |214|Income          |
|divorced                         |25   |32 |Marital_Status  |
|financial problems               |62   |79 |Financial_Status|
|monthly income of $1200 per month|68   |100|Income          |
|financially dependent            |42   |62 |Financial_Status|
|married                          |27   |33 |Marital_Status  |
|immigrant                        |42   |50 |Population_Group|
|high school graduate             |187  |206|Education       |
|construction worker              |215  |233|Employment      |
+---------------------------------+-----+---+----------------+

Model Information

Model Name:	ner_sdoh_income_social_status
Compatibility:	Healthcare NLP 4.4.4+
License:	Licensed
Edition:	Official
Input Labels:	[sentence, token, embeddings]
Output Labels:	[ner]
Language:	en
Size:	3.0 MB
Dependencies:	embeddings_clinical

References

Internal SDOH Project

Benchmarking

           label  precision    recall  f1-score   support
       Education       0.85      0.92      0.88        61
      Employment       0.94      0.96      0.95      2142
Financial_Status       0.89      0.76      0.82       128
          Income       0.97      0.80      0.88        41
  Marital_Status       0.97      0.97      0.97        93
Population_Group       0.91      0.87      0.89        23
       micro-avg       0.94      0.95      0.94      2488
       macro-avg       0.92      0.88      0.90      2488
    weighted-avg       0.94      0.95      0.94      2488

PREVIOUSExtract Health Behaviors and Problems Entities from Social Determinants of Health Texts

NEXTExtract Social Environment Entities from Social Determinants of Health Texts