CoolFace
Datasetpublic

Khyatimirani/pcos-clinical-guideline-101

PCOS Clinical Q&A Dataset (Dr. Thaïs Aliabadi Guideline-Based) 📌 Overview This dataset contains 500+ clinically grounded conversational Q&A pairs simulating interactions between a PCOS patient and a clinically informed assistant. The dataset is derived from the following primary source: Female Hormone Health, PCOS, Endometriosis, Fertility & Breast Cancer | Dr. Thaïs Aliabadi This conversation features Dr. Thaïs Aliabadi (board-certified OB/GYN) explaining… See the full description on the dataset page: https://huggingface.co/datasets/Khyatimirani/pcos-clinical-guideline-101.

sourceHugging Faceapache-2.0updated 6mo agoView on Hugging Face
0likes15downloads
Dataset Card

PCOS Clinical Q&A Dataset (Dr. Thaïs Aliabadi Guideline-Based)

📌 Overview

This dataset contains 500+ clinically grounded conversational Q&A pairs simulating interactions between a PCOS patient and a clinically informed assistant.

The dataset is derived from the following primary source:

Female Hormone Health, PCOS, Endometriosis, Fertility & Breast Cancer | Dr. Thaïs Aliabadi

This conversation features Dr. Thaïs Aliabadi (board-certified OB/GYN) explaining symptoms, diagnosis criteria, root causes, and treatment approaches for PCOS and related conditions.

🎯 Objective

To build a clinically reliable conversational dataset that:

Reflects real-world patient questions Provides clear, medically grounded explanations Includes statistics and prevalence data where available Helps users understand the “why” behind symptoms and treatments Encourages safe escalation to healthcare providers when needed.

🧠 Dataset Design Philosophy

Single source of truth: Dr. Thaïs Aliabadi’s clinical explanations No hallucinated or external medical knowledge added Questions are reverse-engineered from real patient behavior Answers are: Clinically accurate Empathetic but not fluffy Educational and explanatory Safety-aware

📊 Key Medical Grounding (from source)

PCOS affects ~15–20% of women of reproductive age Diagnosis is based on 2 out of 3 criteria: Androgen excess (acne, hair growth, hair thinning) Ovulatory dysfunction (irregular periods) Polycystic ovaries / high AMH PCOS is often underdiagnosed or misdiagnosed despite being highly prevalent Insulin resistance and inflammation are key underlying drivers

📂 Dataset Structure

The dataset is provided in JSONL format:

{ "id": "pcos_001", "section": "Diagnosis", "tags": ["irregular_periods", "androgens"], "messages": [ { "role": "user", "content": "My periods are very irregular. Could this be PCOS?" }, { "role": "assistant", "content": "Irregular periods are one of the key indicators of ovulatory dysfunction, which is one of the three diagnostic criteria for PCOS. Typically, a diagnosis requires meeting at least two of the three criteria, so it would be important to evaluate for other features as well." } ] }

🧩 Topics Covered

PCOS symptoms (cycles, acne, hair growth, weight changes) Diagnosis & clinical criteria Hormonal imbalance & androgens Insulin resistance & inflammation Fertility & ovulation Lifestyle interventions Supplements (evidence-based discussion from source) Mental health considerations Misdiagnosis & delayed diagnosis When to seek medical help

⚠️ Safety & Intended Use

This dataset is for: Education AI assistant training Research The assistant: Does not prescribe medication Does not replace a doctor Encourages clinical consultation when necessary

🚀 Use Cases

Fine-tuning LLMs for women’s health assistants Building PCOS-specific AI companions Clinical conversational AI research Patient education tools Preventive care applications

🧪 Limitations

Based on a single expert source Not a substitute for: Clinical diagnosis Personalized treatment Some clinical nuances may require multi-source validation

🤝 Contributions

We welcome:

Multi-source clinical expansion Localization (Hindi / voice-first) IVF, fertility, and menstrual health extensions

📜 License

(Choose one based on your preference)

CC BY-NC 4.0 (recommended for healthcare datasets) or custom research-use license

💬 Contact

If you are working in:

Women’s health PCOS / fertility Healthcare AI

We’d love to collaborate and build meaningful solutions in this space.

email: khyatimirani77@gmail.com