Hugging Face's logo

Datasets:
BrunoDCDO
/
enatjus_v2

Tasks:
Text Classification
Text Generation
Question Answering
Modalities:
Text
Formats:
parquet
Languages:
Portuguese
Size:
1M - 10M
Tags:
healthcare
legal
medical
technical-notes
judicial
Libraries:
Datasets
Dask
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
enatjus_v2 / README.md
BrunoDCDO's picture
BrunoDCDO
Atualiza README - 2026-01-16
1735fec verified
preview code
|
raw
history blame contribute delete
9.7 kB
metadata 
license: cc-by-4.0
task_categories:
  - text-classification
  - text-generation
  - question-answering
language:
  - pt
tags:
  - healthcare
  - legal
  - medical
  - technical-notes
  - judicial
size_categories:
  - 100K<n<1M

e-NatJus Technical Notes Dataset v2

Dataset Description

This dataset contains structured data extracted from technical notes produced by the e-NatJus system (Núcleo de Apoio Técnico do Judiciário), a Brazilian judicial support program that provides evidence-based technical advice on health technologies in litigation contexts.

The e-NatJus program assists judges in making informed decisions about healthcare-related lawsuits by providing technical assessments of medicines, procedures, and health products requested through the Brazilian judicial system.

Dataset Summary

  • Total Records: 402,236 technical notes
  • ID Range: 86 to 453,325
  • Language: Portuguese (pt-BR)
  • Format: Apache Parquet
  • Encoding: UTF-8
  • Columns: 71 (raw) / 72 (treated)

Available Files

File Description Size
base_enatjus.parquet Raw data with original codes ~454 MB
base_enatjus_tratada.parquet Processed data with readable values ~456 MB

The treated version (base_enatjus_tratada.parquet) includes:

  • All codes converted to human-readable values (e.g., "S" → "Sim", "1" → "Medicamento")
  • origem_tratada: Brazilian state (UF) of the responsible NatJus
  • link_parecer: Direct URL to view the original technical note (replaces arquivo column)
  • HTML tags removed and entities decoded in text fields (txaConclusao, txaEficaciaSeguranca, etc.)

Supported Tasks

  • Healthcare technology assessment analysis
  • Medical-legal text classification
  • Evidence-based medicine research
  • Healthcare policy analysis
  • Drug and procedure recommendation systems
  • Judicial decision support systems

Dataset Structure

Data Fields

The dataset contains 65 fields organized into the following categories:

1. Identification and Metadata

  • titulo (str): Technical note title
  • idNotaTecnica (str): Unique identifier
  • arquivo (str): Source HTML filename
  • status (str): Processing status

2. Patient Data

  • txtIdade (str): Patient age
  • selStaGenero (str): Gender (m/f)
  • txtCidade (str): City/municipality

3. Legal Representative

  • txtNomeAdvogado (str): Attorney name
  • txtNumeroOABAdvogado (str): Bar association number
  • selDefensoriaPublica (str): Public defender/prosecutor indicator

4. Procedural Data

  • selEsfera (str): Judicial sphere (state/federal)
  • txtServentia (str): Court/jurisdiction

5. Clinical Assessment

  • txtCid (str): ICD-10 diagnosis code
  • txtDescAvaliacaoDiagnosticoSemCID (str): Diagnosis description
  • txaMeioDiagRealizado (str): Diagnostic procedures performed

6. Technology Type

  • selTipoTecnologia (str): Technology type
    • "1": Medicine
    • "2": Procedure
    • "3": Product

7. Medicine-Specific Fields (when type = "1")

  • txtDcb (str): Active pharmaceutical ingredient
  • txtDcbComercial (str): Commercial name
  • txtViaAdministracao (str): Route of administration
  • txaPosologia (str): Dosage/posology

8. Procedure-Specific Fields (when type = "2")

  • txtProcedimento (str): Procedure description

9. Product-Specific Fields (when type = "3")

  • txtProduto (str): Health product description

10. Regulatory Status (medicines and products only)

  • selRegistroAnvisa (str): ANVISA registration (S/N)
  • selSituacaoAnvisa (str): Registration status (A=Valid, I=Expired)

11. SUS Availability

  • selDisponivelSus (str): Available in public health system (S/N/B/X)
  • selTabelaTecnologia (str): Incorporation table
    • "R": RENAME (National Essential Medicines)
    • "M": REMUME (Municipal Medicines)
    • "S": SIGTAP (Procedures Table)
    • "C": CIB Deliberation
    • "N": None

12. Cost Information (mainly for medicines)

  • txtLaboratorio (str): Manufacturer
  • txtMarcaComercial (str): Commercial brand
  • txtApresentacao (str): Packaging/presentation
  • txtPrecoFabrica (str): Factory price
  • txtPrecoMaximoGoverno (str): Maximum government price
  • txtPrecoMaximoConsumidor (str): Maximum consumer price

13. Evidence and Technical Foundation

  • txaEficaciaSeguranca (str): Efficacy and safety analysis
  • txaImpactoTecnologia (str): Technology impact assessment
  • selRecomendacaoConitec (str): CONITEC recommendation (F/D/V)
  • selEvidenciaCientifica (str): Scientific evidence indicator (S/N/B)
  • txaReferencia (str): Bibliographic references

14. Conclusion and Opinion

  • selConclusao (str): Final opinion (F=Favorable, N=Unfavorable)
  • txaConclusao (str): Full conclusion text
  • selAlegacaoUrgencia (str): Urgency allegation (S/N)

15. Responsible Parties

  • txtNatResponsavel (str): Technical responsible
  • txtInstituicaoResponsavel (str): Responsible institution
  • selApoioTutoria (str): Tutoring support indicator (S/N)
  • origem_natjus (str): e-NatJus origin/nucleus

16. Status and Dates

  • data_emissao (str): Emission date and time (DD/MM/YYYY HH:MM:SS)
  • status_tecnologia (str): Technology status

17. Attachments

  • anexos_count (int): Number of attachments (0-4)
  • anexos_info (str): JSON with attachment details
  • anexo_filename, anexo2_filename, anexo3_filename, anexo4_filename (str): Individual filenames
  • anexo_hash, anexo2_hash, anexo3_hash, anexo4_hash (str): Download hashes
  • anexo_download_url, anexo2_download_url, anexo3_download_url, anexo4_download_url (str): Download URLs

Data Conventions

The dataset uses the following conventions for missing or non-applicable values:

  • String value: Field filled with actual content
  • "NÃO_PREENCHIDO": Field exists but is empty/blank
  • "NÃO_APLICÁVEL": Field does not apply to the specific technology type
  • None: Parsing error or HTML element not found

Conditional Fields

Some fields are only applicable under certain conditions:

  • selSituacaoAnvisa: Only when selRegistroAnvisa = "S" (has registration)
  • selTabelaTecnologia: Only when selDisponivelSus = "S" (available in SUS)
  • Medicine-specific fields: Only when selTipoTecnologia = "1"
  • Procedure-specific fields: Only when selTipoTecnologia = "2"
  • Product-specific fields: Only when selTipoTecnologia = "3"

Data Loading

Using Polars (Recommended) 

import polars as pl

# Lazy loading (memory efficient)
df = pl.scan_parquet("base_enatjus.parquet")

# Filter medicines only
medicamentos = df.filter(pl.col("selTipoTecnologia") == "1").collect()

# Count by technology type
tipo_counts = df.group_by("selTipoTecnologia").agg(
    pl.count().alias("count")
).collect()

Using Pandas 

import pandas as pd

# Load full dataset
df = pd.read_parquet("base_enatjus.parquet")

# Load with column selection (memory efficient)
df = pd.read_parquet(
    "base_enatjus.parquet",
    columns=["idNotaTecnica", "txtDcb", "selConclusao"]
)

Using DuckDB 

import duckdb

# Query without loading into memory
result = duckdb.query("""
    SELECT selTipoTecnologia, COUNT(*) as total
    FROM 'base_enatjus.parquet'
    GROUP BY selTipoTecnologia
""").to_df()

Data Quality

  • Filtered content: Files with error messages were excluded:
    • "Permissão negada" (Permission denied)
    • "Nota Técnica não encontrada" (Technical note not found)
    • "É necessário estar logado" (Login required)
  • Minimum size: Files under 1000 characters were excluded
  • Encoding: All files processed with UTF-8 encoding
  • Parser: BeautifulSoup with 'html.parser'

Use Cases

  1. Healthcare Technology Assessment: Analyze patterns in clinical evidence and recommendations
  2. Pharmaceutical Research: Study medicine prescriptions, dosages, and availability
  3. Legal Analytics: Understand judicial health litigation patterns
  4. Evidence-Based Medicine: Research correlation between scientific evidence and recommendations
  5. Health Economics: Analyze pricing data and cost-effectiveness
  6. Public Health Policy: Assess SUS availability and incorporation of technologies
  7. NLP Applications: Train models for medical-legal text understanding

Limitations

  • Data represents Brazilian healthcare and legal context
  • Some fields may be incomplete due to original data entry practices
  • Monetary values are stored as strings and may require parsing
  • Attachment files are not included (only metadata)
  • Temporal coverage varies (check data_emissao field)

Citation

If you use this dataset in your research, please cite:

@dataset{enatjus_v2_2025,
  title={e-NatJus Technical Notes Dataset v2},
  author={Brazilian Judicial Technical Support Program},
  year={2025},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/BrunoDCDO/enatjus_v2}
}

License

This dataset is released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.

You are free to:

  • Share — copy and redistribute the material
  • Adapt — remix, transform, and build upon the material

Under the following terms:

  • Attribution — You must give appropriate credit

Contact

For questions or issues regarding this dataset, please open an issue on the Hugging Face dataset page.

Additional Resources