| --- |
| license: cc-by-4.0 |
| task_categories: |
| - tabular-classification |
| - tabular-regression |
| language: |
| - en |
| tags: |
| - healthcare |
| - supply-chain |
| - laboratory |
| - reagent |
| - diagnostics |
| - GeneXpert |
| - molecular |
| - ASLM |
| - sub-saharan-africa |
| - lmic |
| pretty_name: "Laboratory Reagent Supply (Reagent Availability, Equipment, Diagnostic Access, Stockouts)" |
| size_categories: |
| - 10K<n<100K |
| configs: |
| - config_name: reference_laboratory |
| data_files: data/lab_reference_laboratory.csv |
| - config_name: district_laboratory |
| data_files: data/lab_district_laboratory.csv |
| default: true |
| - config_name: health_centre_lab |
| data_files: data/lab_health_centre_lab.csv |
| --- |
| |
| # Laboratory Reagent Supply Dataset |
|
|
| ## Abstract |
|
|
| This dataset provides **30,000 simulated facility-level observations** (10,000 per scenario) of laboratory reagent and consumable availability across three tiers of the laboratory network in sub-Saharan Africa. Each record represents one reagent assessed at one laboratory during one quarterly period. The dataset captures 35+ variables spanning reagent type, department, availability, stockout metrics, equipment functionality, tests missed, sample referral, and supply chain performance. Three scenarios: reference laboratory (80% availability), district laboratory (47%), health centre (22%). |
|
|
| **This dataset is entirely simulated. It must not be used for clinical or procurement decisions.** |
|
|
| ## 1. Introduction |
|
|
| ### 1.1 The Diagnostic Access Crisis |
|
|
| The Lancet Commission on Diagnostics (2021) estimated that 47% of the global population — predominantly in LMICs — lacks access to essential diagnostics, with only 1% of health spending allocated to diagnostics. In sub-Saharan Africa, laboratory services are among the most neglected components of health systems, despite their critical role in disease surveillance, treatment monitoring, and antimicrobial stewardship. |
|
|
| ### 1.2 Reagent Supply Challenges |
|
|
| Laboratory reagent supply in SSA faces unique challenges: cold chain requirements for many reagents, single-source dependency for proprietary platforms (GeneXpert, Abbott), import clearance delays, and misalignment between equipment procurement and reagent supply contracts. The WHO Service Availability and Readiness Assessment (SARA) consistently documents low laboratory readiness, with reagent stockouts as a primary barrier. |
|
|
| ### 1.3 Molecular Diagnostics |
|
|
| The GeneXpert platform has been widely deployed for TB diagnosis and HIV viral load monitoring through PEPFAR. However, cartridge supply is frequently disrupted, with facilities reporting stockouts of 30-60% for molecular reagents. The African Society for Laboratory Medicine (ASLM) has highlighted the need for integrated reagent supply chain management. |
|
|
| ### 1.4 Rationale |
|
|
| This dataset integrates reagent availability across departments (biochemistry, haematology, microbiology, molecular, serology, POC), equipment functionality, and diagnostic service impact for laboratory supply chain modelling and health systems research. |
|
|
| ## 2. Methodology |
|
|
| ### 2.1 Parameterization |
|
|
| | Parameter | Reference Lab | District Lab | Health Centre | Source | |
| | --- | --- | --- | --- | --- | |
| | Reagent availability | 82% | 58% | 32% | WHO SARA | |
| | Molecular availability | 70% | 40% | 2% | PEPFAR lab network | |
| | Equipment functional | 80% | 55% | 25% | ASLM assessments | |
| | Stocked out 6m | 26% | 52% | 75% | Multi-country data | |
| | Expired reagent rate | 8% | 18% | 30% | WHO SARA | |
| | Order fill rate | 78% | 52% | 25% | Lab supply data | |
|
|
| ### 2.2 Reagent Selection |
|
|
| 18 reagents across 9 departments: biochemistry (glucose, chemistry analyzer), haematology (Hb, CBC), urinalysis, blood bank (grouping antisera), serology (RPR, HBV, HCV), molecular (GeneXpert TB/VL/EID), parasitology (Giemsa), microbiology (Gram stain, blood culture, Mueller-Hinton), immunology (CD4), POC (pregnancy test). |
|
|
| ### 2.3 Scenario Design |
|
|
| **Scenario A — Reference Laboratory**: National/regional reference lab with molecular platforms, culture capability, chemistry analyzers, cold storage, 8+ lab scientists. |
|
|
| **Scenario B — District Laboratory**: GeneXpert for TB, chemistry analyzer, no culture capability, limited cold storage, 3 staff. |
|
|
| **Scenario C — Health Centre Lab**: POC tests only (RDTs, dipsticks, Hb), no molecular or culture, no cold storage, 1 lab assistant. |
|
|
| ## 3. Schema |
|
|
| | Column | Type | Description | |
| | --- | --- | --- | |
| | facility_level | categorical | reference_lab / district_lab / health_centre | |
| | reagent_name | categorical | 18 laboratory reagents | |
| | department | categorical | biochemistry / haematology / microbiology / molecular / serology / parasitology / immunology / POC / blood_bank / urinalysis | |
| | format | categorical | reagent type format | |
| | cold_chain_required | binary | Requires cold storage | |
| | available_on_survey_day | binary | Reagent available | |
| | stocked_out_in_last_6m | binary | Any stockout in 6 months | |
| | stockout_days_last_6m | int | Total stockout days | |
| | stockout_cause | categorical | 11 root cause categories | |
| | equipment_functional | binary | Lab equipment operational | |
| | tests_performed_month | int | Tests done this period | |
| | tests_not_done_no_reagent | int | Tests not done due to stockout | |
| | sample_referred_out | binary | Sample sent to higher lab | |
| | turnaround_time_days | float | Result turnaround time | |
| | expired_reagent_found | binary | Expired reagent on shelf | |
| | months_of_stock | float | Stock in months | |
| | order_fill_rate_pct | float | Order fulfilment rate | |
| |
| ## 4. Validation |
| |
| <p align="center"> |
| <img src="validation_report.png" alt="Validation Report" width="100%"> |
| </p> |
| |
| ## 5. Usage |
| |
| ```python |
| from datasets import load_dataset |
|
|
| dataset = load_dataset( |
| "electricsheepafrica/laboratory-reagent-supply", |
| "district_laboratory" |
| ) |
| df = dataset["train"].to_pandas() |
| |
| # Which departments are most affected by stockouts? |
| print(df.groupby('department')['available_on_survey_day'].mean().sort_values()) |
| ``` |
| |
| ## 6. Limitations |
| |
| - **Simulated**: Not from real laboratory information systems. |
| - **No platform lock-in**: Proprietary reagent-equipment dependencies simplified. |
| - **No quality assurance**: External quality assessment participation not modelled. |
| |
| ## 7. References |
| |
| 1. Lancet Commission on Diagnostics (2021). 47% lack access to diagnostics. |
| 2. WHO SARA. Service Availability and Readiness Assessment. Laboratory indicators. |
| 3. ASLM. African Society for Laboratory Medicine. Lab strengthening frameworks. |
| 4. PEPFAR. Laboratory network and GeneXpert reagent supply data. |
| |
| ## Citation |
| |
| ```bibtex |
| @dataset{esa_lab_reagent_2025, |
| title = {Laboratory Reagent Supply Dataset}, |
| author = {{Electric Sheep Africa}}, |
| year = {2025}, |
| publisher = {Hugging Face}, |
| url = {https://huggingface.co/datasets/electricsheepafrica/laboratory-reagent-supply}, |
| note = {Simulated dataset. Not for clinical or procurement decisions.} |
| } |
| ``` |
| |
| ## License |
| |
| [CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/) |
| |
