---
license: cc-by-4.0
task_categories:
  - tabular-classification
  - tabular-regression
language:
  - en
tags:
  - healthcare
  - supply-chain
  - HIV
  - ARV
  - antiretroviral
  - test-kit
  - PEPFAR
  - viral-load
  - ART
  - sub-saharan-africa
  - lmic
pretty_name: "HIV Test Kit & ARV Supply (Commodity Availability, Stockouts, ART Interruption, Patient Impact)"
size_categories:
  - 10K<n<100K
configs:
  - config_name: pepfar_supported_urban
    data_files: data/hiv_supply_pepfar_supported_urban.csv
  - config_name: district_hospital_art
    data_files: data/hiv_supply_district_hospital_art.csv
    default: true
  - config_name: rural_health_centre_art
    data_files: data/hiv_supply_rural_health_centre_art.csv
---

# HIV Test Kit & ARV Supply Dataset

## Abstract

This dataset provides **30,000 simulated facility-level observations** (10,000 per scenario) of HIV commodity availability across three tiers of ART service delivery in sub-Saharan Africa. Each record represents one HIV commodity (ARVs, test kits, viral load reagents, OI prophylaxis) assessed at one facility during one quarterly period. The dataset captures 40+ variables spanning facility characteristics, commodity type, day-of-survey availability, stockout duration and causes, supply chain metrics, and patient impact including ART interruption, regimen substitution, and missed testing. Three scenarios model the HIV supply chain gradient: PEPFAR-supported urban ART clinic (89% availability), district hospital (65%), and rural health centre (30%).

**This dataset is entirely simulated. It must not be used for clinical decisions or programme management.**

## 1. Introduction

### 1.1 The HIV Commodity Supply Chain

Sub-Saharan Africa bears the largest HIV burden globally, with over 25 million people living with HIV and approximately 21 million on antiretroviral therapy (ART). Sustaining this treatment cascade requires an unbroken supply of ARV medicines, HIV rapid diagnostic test kits, viral load monitoring reagents, and opportunistic infection prophylaxis. The supply chain is predominantly funded by PEPFAR and the Global Fund, with USAID's Global Health Supply Chain Program (GHSC-PSM) serving as the primary procurement and distribution mechanism.

### 1.2 ARV Availability: Progress and Persistent Gaps

GHSC-PSM has delivered enough ARVs to provide over 20.5 million patient-years of treatment, including 79 million bottles of TLD (tenofovir/lamivudine/dolutegravir) — the WHO-recommended first-line regimen (GHSC-PSM, 2023). TLD availability at PEPFAR-supported sites increased from 84% to 100% between June 2020 and June 2021, and has remained between 98% and 100% since (GHSC-PSM, "Crisis to Confidence"). The Global Fund reduced TLD pricing to below $45 per patient per year through competitive tenders (Global Fund, 2023).

However, these achievements mask significant gaps. UNAIDS reported in 2025 that only 39% of reporting countries indicated low-to-minimal risk of ARV stockouts, meaning 61% face moderate-to-high risk — particularly countries dependent on donor funding transitions (UNAIDS, 2025). Pediatric ARV availability remains substantially lower than adult formulations, with nevirapine suspension and other pediatric formulations showing persistent shortages (GHSC-PSM; Nation Africa, 2024).

### 1.3 Test Kit and Reagent Supply

HIV rapid diagnostic test (RDT) kits and viral load reagents face distinct supply chain challenges. Test kit stockouts directly impede case finding — the entry point to the treatment cascade. In rural Mozambique, monitoring of pharmacy and test kit stocks across PEPFAR-supported facilities documented frequent stockouts in Zambézia Province, with rural facilities disproportionately affected (PMC, 2020; PMC7232670). Viral load reagent availability is particularly constrained at lower facility levels, where laboratory infrastructure is absent and sample referral networks are weak.

### 1.4 Patient Impact of Stockouts

ARV stockouts force treatment interruptions that undermine viral suppression, promote drug resistance, and increase mortality. Regimen substitutions — switching patients to available but suboptimal regimens — carry risks of adverse effects and reduced efficacy. Test kit stockouts mean undiagnosed infections continue to drive transmission. Viral load monitoring gaps mean treatment failure goes undetected.

### 1.5 Rationale

Despite extensive PEPFAR and Global Fund reporting, no open, standardised dataset integrates facility-level HIV commodity availability, stockout dynamics, supply chain metrics, and patient impact indicators. This dataset enables supply chain analysts and researchers to develop predictive models, benchmark facility performance, and model intervention strategies.

## 2. Methodology

### 2.1 Epidemiological Parameterization

| Parameter | PEPFAR Urban | District Hospital | Rural HC | Source |
| --- | --- | --- | --- | --- |
| ARV availability | 92% | 72% | 45% | GHSC-PSM: 98-100% TLD at PEPFAR sites |
| Test kit availability | 88% | 65% | 35% | PMC7232670: rural Mozambique |
| VL reagent availability | 70% | 35% | 5% | Lab infrastructure constraints |
| Pediatric ARV availability | 75% | 50% | 25% | GHSC-PSM: lower than adult |
| Stocked out in 6m | 17% | 54% | 95% | UNAIDS: 61% countries moderate-high risk |
| Mean stockout days | 5 | 15 | 31 | Literature range |
| ART interruption rate | 0.5% | 5% | 18% | Clinical impact data |
| Order fill rate | 84% | 60% | 35% | GHSC-PSM end-use verification |

### 2.2 HIV Commodity Selection

18 commodities across 8 categories reflecting the complete HIV supply chain:

| Category | Commodities |
| --- | --- |
| ARV first-line | TLD (adult), TLE (adult) |
| ARV component | DTG 50mg |
| ARV second-line | ATV/r (adult) |
| ARV paediatric | LPV/r granules, Nevirapine suspension, ABC/3TC dispersible, DTG 10mg dispersible |
| ARV PMTCT | AZT syrup |
| Test kits | HIV RDT (screening), HIV RDT (confirmatory), HIV self-test |
| Lab reagents | Viral load cartridge, EID cartridge, CD4 cartridge |
| OI prophylaxis/TPT | Cotrimoxazole, Fluconazole, Isoniazid |

### 2.3 Scenario Design

**Scenario A — PEPFAR-Supported Urban ART Clinic**: Electronic LMIS, dedicated HIV pharmacist, multi-month dispensing (MMD), direct GHSC-PSM supply, 3,000+ active patients. Analogous to high-volume ART sites in Kenya, Tanzania, Uganda, South Africa.

**Scenario B — District Hospital ART Department**: Paper LMIS, clinical officer managing ART, quarterly supply from district pharmacy, 800 active patients. Analogous to district ART sites in Malawi, Mozambique, Rwanda, Zambia.

**Scenario C — Rural Health Centre ART**: Nurse-led ART, no LMIS, irregular supply, limited test kits, no on-site viral load, 150 active patients. Analogous to primary ART sites in DRC, Niger, South Sudan, rural Mozambique.

## 3. Schema

### 3.1 Facility & Programme

| Column | Type | Description |
| --- | --- | --- |
| facility_level | categorical | urban_ART_clinic / district_hospital / rural_health_centre |
| pepfar_supported | binary | PEPFAR funding support |
| global_fund_supported | binary | Global Fund support |
| has_hiv_pharmacist | binary | Dedicated HIV pharmacist |
| lmis_type | categorical | electronic / paper / none |
| mmd_available | binary | Multi-month dispensing offered |
| art_patients_active | int | Active patients on ART |
| new_art_initiations_quarter | int | New initiations this quarter |

### 3.2 Commodity

| Column | Type | Description |
| --- | --- | --- |
| commodity_name | categorical | 18 HIV commodities |
| commodity_category | categorical | ARV / test_kit / lab_reagent / OI / TB_preventive |
| formulation | categorical | tablet_FDC / tablet / suspension / granules / rapid_test / self_test / reagent_cartridge |
| pepfar_funded_commodity | binary | Commodity procured via PEPFAR |
| unit_cost_usd | float | Approximate unit cost |

### 3.3 Availability & Stockout

| Column | Type | Description |
| --- | --- | --- |
| available_on_survey_day | binary | Commodity present and usable |
| stocked_out_in_last_6m | binary | Any stockout in 6 months |
| stockout_days_last_6m | int | Total stockout days |
| stockout_episodes | int | Number of stockout episodes |
| stockout_cause | categorical | 12 root cause categories |

### 3.4 Supply Chain

| Column | Type | Description |
| --- | --- | --- |
| months_of_stock_on_hand | float | Current stock in months |
| last_resupply_days_ago | int | Days since last delivery |
| order_fill_rate_pct | float | Percentage of order fulfilled |
| emergency_order_placed | binary | Emergency order submitted |

### 3.5 Patient Impact

| Column | Type | Description |
| --- | --- | --- |
| patients_turned_away | int | Patients turned away due to stockout |
| art_interruption_due_to_stockout | binary | ART treatment interrupted |
| regimen_substitution | binary | Switched to alternative regimen |
| testing_not_done_no_kits | binary | HIV testing not performed |
| viral_load_not_done | binary | VL monitoring skipped |

### 3.6 Data Quality

| Column | Type | Description |
| --- | --- | --- |
| stock_card_up_to_date | binary | Stock records current |
| report_submitted_to_district | binary | Quarterly report submitted |
| report_timely | binary | Report on time |

## 4. Validation

<p align="center">
  <img src="validation_report.png" alt="Validation Report" width="100%">
</p>

### 4.1 Key Validation Results

| Metric | PEPFAR Urban | District | Rural | Literature |
| --- | --- | --- | --- | --- |
| Availability | 89% | 65% | 30% | 98-100% PEPFAR [2]; gaps elsewhere [3] |
| Stocked out 6m | 17% | 54% | 95% | 61% countries at risk [3] |
| Stockout duration | 5d | 15d | 31d | Literature range |
| ART interruption | 0.5% | 5% | 18% | Clinical impact |
| Order fill rate | 84% | 60% | 35% | GHSC-PSM data |

## 5. Usage

```python
from datasets import load_dataset

dataset = load_dataset(
    "electricsheepafrica/hiv-testkit-arv-supply",
    "district_hospital_art"
)
df = dataset["train"].to_pandas()

# Which commodities are most at risk?
risk = df.groupby('commodity_name').agg({
    'available_on_survey_day': 'mean',
    'art_interruption_due_to_stockout': 'mean'
}).sort_values('available_on_survey_day')
print(risk.head(10))
```

## 6. Limitations

- **Simulated**: Not from real PEPFAR or Global Fund supply chain data.
- **No donor transition modelling**: The dataset does not simulate the impact of PEPFAR funding changes.
- **No drug resistance**: ART interruptions do not model downstream resistance emergence.
- **No longitudinal tracking**: Records are cross-sectional, not patient-level treatment histories.
- **Simplified costing**: Unit costs are approximate global reference prices.

## 7. References

1. GHSC-PSM / USAID (2023). FY23 Q1 IDIQ Report. 20.5M patient-years ARV delivered.
2. GHSC-PSM. From Crisis to Confidence. TLD availability 98-100%.
3. UNAIDS (2025). HIV commodity availability fact sheet. 39% low risk; 61% moderate-high risk.
4. Global Fund (2023). TLD price reduced to <$45/year.
5. PMC (2020). Monitoring pharmacy and test kit stocks in rural Mozambique. PMC7232670.
6. Nation Africa (2024). Kenya among six countries on brink of running out of HIV drugs.

## Citation

```bibtex
@dataset{esa_hiv_testkit_arv_2025,
  title   = {HIV Test Kit and ARV Supply Dataset: Commodity Availability,
             Stockouts, ART Interruption, and Patient Impact Across Three
             Tiers of ART Service Delivery in Sub-Saharan Africa},
  author  = {{Electric Sheep Africa}},
  year    = {2025},
  publisher = {Hugging Face},
  url     = {https://huggingface.co/datasets/electricsheepafrica/hiv-testkit-arv-supply},
  note    = {Simulated dataset. Not for clinical or programme decisions.}
}
```

## License

[CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)