electricsheepafrica/surgical-outcomes-perioperative

Surgical Capacity & Perioperative Outcomes Dataset

Abstract

This dataset provides 30,000 simulated surgical records (10,000 per scenario) from hospitals in sub-Saharan Africa. Each record contains 40+ variables including patient demographics, ASA grade, procedure type, urgency, anaesthesia type/provider, WHO surgical safety checklist use, pulse oximetry availability, blood loss, surgical site infection, complications, and mortality. Three facility levels: tertiary hospital (4.2% mortality), district hospital (10.6%), and rural surgical camp (21.4%).

1. Introduction

Five billion people lack access to safe, affordable surgery (Meara et al., Lancet 2015). The ASOS study across 25 African countries found 2.1% in-hospital mortality — twice the global average — despite a low-risk patient profile (Biccard et al., Lancet 2018). Complications occurred in 18.2%, with infection the most common (10.2%). SSA has only 0.7 SAO providers per 100,000 population vs the target of 20 (Ng-Kamstra et al., 2016). The ASOS-Paeds study (2024) found 2.3% paediatric surgical mortality in Africa — 11-fold higher than HICs. GlobalSurg reported SSI rates of 7.2% overall in Africa, reaching 23% for GI surgery in low-HDI countries.

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

2. Methodology

2.1 Parameterization

Parameter	Value	Source
Perioperative mortality (Africa)	2.1% overall	Biccard et al., ASOS Lancet 2018
Paediatric surgical mortality	2.3% (11x HICs)	ASOS-Paeds, Lancet 2024
Complication rate	18.2%	Biccard et al., ASOS Lancet 2018
SSI rate (Africa)	7.2–23%	GlobalSurg, Lancet Infect Dis 2018
Emergency proportion	57%	ASOS 2018
SAO workforce density	0.7/100K	Ng-Kamstra et al., World J Surg 2016
Checklist mortality reduction	47%	Haynes et al., NEJM 2009
Anaesthesia providers <5/100K	>70 countries	Law et al., WFSA/Anesth Analg 2024

2.2 Scenario Design

Scenario	Facility	Checklist	Pulse Ox	SSI	Mortality
Tertiary hospital	Teaching hospital	70%	95%	15%	4.2%
District hospital	District, 1-2 ORs	30%	50%	32%	10.6%
Rural surgical camp	Intermittent surgery	11%	20%	48%	21.4%

3. Schema

Column	Type	Description
id	int	Unique identifier
age_years	int	Patient age
sex	categorical	M / F
asa_grade	int	ASA physical status (1–5)
hiv_positive	binary	HIV status
diabetes	binary	Diabetes
hypertension	binary	Hypertension
anaemia	binary	Anaemia
urgency	categorical	emergency / elective
procedure	categorical	caesarean_section / hernia_repair / appendicectomy / laparotomy / fracture_fixation / etc.
specialty	categorical	general_surgery / obstetrics_gynaecology / orthopaedics / plastics_burns
indication	categorical	emergency / elective
pulse_oximetry	binary	Pulse oximetry available
checklist_used	binary	WHO surgical safety checklist used
blood_available	binary	Blood products available
anaesthesia_provider	categorical	physician_anaesthetist / nurse_anaesthetist / clinical_officer / medical_officer
anaesthesia_type	categorical	general / spinal / local / ketamine / regional_block / epidural
surgical_duration_min	int	Operative time (minutes)
estimated_blood_loss_ml	int	Estimated blood loss
blood_transfusion	binary	Received transfusion
surgical_site_infection	binary	SSI developed
anaesthesia_complication	binary	Anaesthesia complication
postop_haemorrhage	binary	Postoperative haemorrhage
wound_dehiscence	binary	Wound dehiscence
pneumonia	binary	Postoperative pneumonia
vte	binary	Venous thromboembolism
any_complication	binary	Any complication
outcome	categorical	survived / died
icu_admission	binary	ICU admission
length_of_stay_days	int	Hospital stay
reoperation	binary	Reoperation required
prophylactic_antibiotics	binary	Prophylactic antibiotics given

4. Validation

Key validation checks:

• Mortality gradient: Tertiary 4.2% → District 10.6% → Rural 21.4% ✓
• Emergency > elective mortality: 2–3× higher, consistent with ASOS/GlobalSurg ✓
• SSI gradient: 15% → 32% → 48%, reflecting checklist/infrastructure gap ✓
• Checklist protective: Lower mortality where checklist used ✓
• ASA dose-response: Mortality increases steeply with ASA grade ✓
• Caesarean most common: Consistent with ASOS (33% CS) ✓

5. Usage

from datasets import load_dataset
dataset = load_dataset("electricsheepafrica/surgical-outcomes-perioperative", "district_hospital")
df = dataset["train"].to_pandas()

python generate_dataset.py --all-scenarios --n 10000 --seed 42

6. Limitations

• Simulated: Not derived from real surgical registries.
• No imaging/lab: No preoperative imaging or lab values.
• Simplified SSI: Binary, not classified by depth (superficial/deep/organ-space).
• No cost data: No surgical/anaesthesia costs or catastrophic expenditure.
• No time-to-OR: Delay from presentation to surgery not modelled.

7. References

1. Meara JG, et al. (2015). Lancet Commission on Global Surgery. Lancet, 386(9993):569–624.
2. Biccard BM, et al. (2018). ASOS: perioperative outcomes in Africa. Lancet, 391(10130):1589–1598.
3. GlobalSurg Collaborative (2018). SSI after GI surgery. Lancet Infect Dis, 18(5):516–525.
4. Ng-Kamstra JS, et al. (2016). Surgical workforce density SSA. World J Surg, 40(6):1369–1376.
5. Haynes AB, et al. (2009). WHO surgical safety checklist. NEJM, 360(5):491–499.
6. Dare AJ, et al. (2014). Surgical conditions mortality. Lancet, 383(Suppl 2):S3.
7. Prin M, et al. (2018). Anaesthesia in SSA. Anesth Analg, 126(4):1.
8. WHO (2018). Safe Surgery Saves Lives.
9. ASOS-Paeds Investigators (2024). Paediatric surgical outcomes in Africa. Lancet, 403(10435):1441–1451.
10. Law TJ, et al. (2024). Global Anesthesia Workforce Survey. Anesth Analg, March 2024.

Citation

@dataset{esa_surgical_2025,
  title={Surgical Capacity and Perioperative Outcomes Dataset},
  author={Electric Sheep Africa},
  year={2025},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/electricsheepafrica/surgical-outcomes-perioperative}
}

License

CC-BY-4.0