electricsheepafrica/rabies-post-exposure

Rabies & Post-Exposure Prophylaxis Dataset

Abstract

This dataset provides 30,000 simulated animal bite/rabies exposure records (10,000 per scenario) from sub-Saharan Africa. Each record contains 45+ variables including exposure type, wound care, PEP access, vaccine completion, RIG, and outcomes. Three settings: urban referral (2.9% rabies mortality), district hospital (4.1%), and rural health centre (9.9%).

1. Introduction

Dogs are responsible for 99% of human rabies cases. Rabies is invariably fatal once symptoms appear but is preventable through timely PEP. SSA bears a high burden due to limited vaccine access, stock-outs, cost barriers, and poor wound care. WHO recommends the intradermal route for PEP. RIG is critical for Category III exposures but rarely available in SSA.

This dataset is entirely simulated. It must not be used for clinical decision-making.

2. Methodology

2.1 Parameterization

Parameter	Value	Source
Dog-transmitted	99%	WHO 2024
Fatal once symptomatic	100%	WHO
PEP preventable	Yes	CDC 2024
RIG access SSA	Very limited	PMC 2019
ID route recommended	Yes	WHO 2024

2.2 Scenario Design

Scenario	Vaccine	RIG	Cold Chain	PEP Complete	Mortality
Urban referral	Yes	Yes	Yes	58%	2.9%
District hospital	Yes	No	Yes	39%	4.1%
Rural HC	No	No	No	1%	9.9%

3. Schema

Column	Type	Description
id	int	Unique identifier
age	int	Patient age
animal_type	categorical	dog / cat / bat / monkey / jackal
who_category	categorical	I / II / III
bite_site	categorical	lower_limb / upper_limb / head_face / hand
wound_washed_soap_water	binary	Wound washed
pep_initiated	binary	PEP started
pep_completed	binary	PEP completed
rig_given	binary	RIG administered
rig_needed_not_given	binary	RIG gap
rabies_developed	binary	Rabies developed
died	binary	Died

4. Validation

Key validation checks:

• PEP completion: 58% → 39% → 1% ✓
• Rabies mortality: 2.9% → 4.1% → 9.9% ✓
• Dog: 80% of exposures ✓
• Category III: ~55% of exposures ✓
• RIG gap: massive outside urban ✓
• Wound washing lower in rural ✓

5. Usage

from datasets import load_dataset
dataset = load_dataset("electricsheepafrica/rabies-post-exposure", "district_hospital")
df = dataset["train"].to_pandas()

6. Limitations

• Simulated: Not from real bite registries.
• No lab confirmation: No DFAT/PCR.
• Simplified: No detailed vaccine schedules.
• No dog population data: No canine vaccination coverage.

7. References

1. WHO (2024). Rabies fact sheet.
2. PMC (2019). PEP vaccine access in LMICs.
3. PMC (2018). WHO position on rabies immunization.
4. WHO (2024). Intradermal PEP recommendations.
5. CDC (2024). Rabies PEP guidance.

Citation

@dataset{esa_rabies_pep_2025,
  title={Rabies and Post-Exposure Prophylaxis Dataset},
  author={Electric Sheep Africa},
  year={2025},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/electricsheepafrica/rabies-post-exposure}
}

License

CC-BY-4.0