EMS-MCQA (subset)

EMS-MCQA is a multiple-choice question answering (MCQA) dataset focused on Emergency Medical Services (EMS) knowledge. See more on our project page.

This repo contains an open-source subset of the full dataset, provided as JSON files.

---

Dataset Summary

Each record is a JSON object with:

• question (str) – question stem
• choices (list[str]) – options (letters like a., b., … are kept)
• answer (str) or (list[str]) – correct option label ("a" | "b" | ["a", "c"])
• explanation (str, optional) – short rationale
• link (str, optional) – source URL
• level (list[str]) – EMS certification level(s), e.g., emr, emt, aemt, paramedic, NA
• category (list[str]) – topical tags (e.g., airway_respiration_and_ventilation, anatomy, assessment, cardiology_and_resuscitation, ems_operations, medical_and_obstetrics_gynecology, pediatrics, pharmacology, trauma)

---

Files & Counts

• train_open.json — 13,021 items
• val_open.json — 1,860 items
• test_open.json — 3,721 items
• MCQA_open_final.json — 18,602 items (entire open-source collection)
• train_open_woNA.json — training split with items where level != "NA"

Use train_open_woNA.json when you want to predict certification/subject without unlabeled levels.

Citation

If you use EMSQA in your work, please cite:

@misc{ge2025expertguidedpromptingretrievalaugmentedgeneration,
  title         = {Expert-Guided Prompting and Retrieval-Augmented Generation for Emergency Medical Service Question Answering},
  author        = {Xueren Ge and Sahil Murtaza and Anthony Cortez and Homa Alemzadeh},
  year          = {2025},
  eprint        = {2511.10900},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  url           = {https://arxiv.org/abs/2511.10900},
}

---

Loading with 🤗 Datasets

from datasets import load_dataset

# Load predefined splits from this repo
ds = load_dataset("Xueren/EMS-MCQA")       # returns a DatasetDict with train/validation/test
print("train:", ds["train"][0])

# Or load the full open-source collection as a single split
full_data   = load_dataset("Xueren/EMS-MCQA", split="full_data")
print("full_data:", full_data[0])

# Or load the train without NA certification label
train_woNA = load_dataset("Xueren/EMS-MCQA", split="train_woNA")
print("train_woNA:", train_woNA[0])