---
license: apache-2.0
language:
  - en
tags:
  - medical
  - clinical
  - emergency-department
  - triage
  - tabular
  - benchmark
size_categories:
  - 100K<n<1M
pretty_name: Hospital Triage and Patient History Data
---

# Hospital Triage and Patient History Data

Tabular dataset of emergency-department triage records and patient history,
suitable for hospital-admission prediction and clinical-tabular LLM
benchmarking.

## Source

This is a **re-hosted** copy of the dataset released by Hong, Haimovich, and
Taylor (Yale) at <https://github.com/yaleemmlc/admissionprediction>. The data
has been **losslessly converted** from the original R `.RData` (via an
intermediate `.feather`) to Apache Parquet with zstd compression so that the
Hugging Face dataset viewer recognizes it. No rows, columns, or values were
modified.

## Shape

- **Rows:** 560,486 emergency-department visits
- **Columns:** 972 variables (triage vitals, chief complaint, medications,
  prior diagnoses, imaging history, labs, demographics)

## Usage

```python
from datasets import load_dataset

ds = load_dataset("kondratevakate/hospital-triage-and-patient-history-data", split="train")
print(ds.num_rows, ds.num_columns)
```

## Citation

If you use this dataset, please cite the original paper:

```bibtex
@article{hong2018predicting,
  title   = {Predicting hospital admission at emergency department triage using machine learning},
  author  = {Hong, Woo Suk and Haimovich, Adrian Daniel and Taylor, Richard Andrew},
  journal = {PLOS ONE},
  volume  = {13},
  number  = {7},
  pages   = {e0201016},
  year    = {2018},
  doi     = {10.1371/journal.pone.0201016}
}
```

## License

Apache-2.0 (as set on the original re-host on this repo). The upstream Yale
release does not specify a separate data license; the authors made the
de-identified dataset publicly available alongside the paper.

## Ethical considerations

Records are de-identified per the original release. Downstream users remain
responsible for compliance with applicable health-data regulations in their
jurisdiction (e.g., HIPAA in the US, GDPR Article 9 in the EU, DHA HIPS in
the UAE) and for obtaining any required institutional approvals (IRB) before
using the data in research.