# HIV Treatment Cascade and Drug Resistance Dataset for Nigeria

## Abstract

This dataset contains synthetic but evidence-based data simulating HIV treatment cascade indicators and drug resistance patterns in Nigeria from 2018-2025. The dataset reflects Nigeria's progress toward the UNAIDS 95-95-95 targets (87% know status, 98% on treatment, 95% virally suppressed as of 2025) and includes detailed information on HIV drug resistance mutations (DRMs) among key populations, with an overall prevalence of 16.7% and higher rates among people who inject drugs (21.6%) and in the North Central zone (25.8%). The dataset supports analysis of HIV program performance, treatment outcomes, and resistance surveillance.

## Introduction

### Background
Nigeria bears the second-largest HIV burden globally, with an estimated 1.9 million people living with HIV (PLHIV) and a national prevalence of 1.3% among adults aged 15-49 years. Despite significant progress in HIV treatment scale-up, challenges remain in achieving epidemic control, particularly regarding HIV drug resistance (HIVDR) which threatens the long-term effectiveness of antiretroviral therapy (ART).

### Problem Statement
While Nigeria has made progress toward the 95-95-95 targets, gaps remain in treatment coverage, viral suppression, and resistance monitoring. Understanding the interplay between treatment cascade indicators and emerging drug resistance patterns is critical for optimizing HIV program strategies, especially among key populations who disproportionately bear the burden of HIVDR.

### Objectives
This dataset aims to:
1. Simulate realistic HIV treatment cascade progression from diagnosis to viral suppression
2. Model HIV drug resistance emergence based on evidence from recent Nigerian studies
3. Enable analysis of factors associated with treatment success and failure
4. Support research on HIVDR surveillance and prevention strategies
5. Facilitate evaluation of program interventions across different burden scenarios

## Methodology

### Parameter Evidence Table
The dataset parameters are grounded in recent Nigerian HIV research and program data:

| Parameter | Value | Source | DOI/Link |
|-----------|-------|--------|----------|
| HIV Prevalence (Adults 15-49) | 1.3% | NACA Nigeria Prevalence Rate | [NACA Nigeria](https://naca.gov.ng) |
| Estimated PLHIV | 1.9 million | NACA Nigeria Prevalence Rate | [NACA Nigeria](https://naca.gov.ng) |
| 95-95-95 Progress (2025) | 87% know status, 98% on treatment, 95% virally suppressed | NACA press conference, Nov 2025 | [NACA Press Release](https://naca.gov.ng/news) |
| New Infections Decline (2025) | 46% drop | The Sun Nigeria, Nov 2025 | [The Sun Nigeria](https://sunnewsonline.com) |
| Transmission Risk Reduction | 95% of Nigerians on HIV treatment cannot transmit virus | Guardian Nigeria, Nov 2025 | [Guardian Nigeria](https://guardian.ng) |
| HIV Drug Resistance Prevalence | 16.7% of key populations have ≥1 DRM | PubMed PMID: 40931993 | DOI: 10.1097/QAI.0000000000003763 |
| DRM Prevalence in PWID | 21.6% | PubMed PMID: 40931993 | DOI: 10.1097/QAI.0000000000003763 |
| DRM Prevalence in North Central Zone | 25.8% | PubMed PMID: 40931993 | DOI: 10.1097/QAI.0000000000003763 |
| Common DRMs | K103N, M41L, M184V | PubMed PMID: 40931993 | DOI: 10.1097/QAI.0000000000003763 |

### DAG-Based Sampling Approach
The dataset generation employs a directed acyclic graph (DAG) approach to ensure logical consistency across the HIV treatment cascade:

1. **Geographic Assignment**: States → LGAs → Zones (based on Nigeria's administrative structure)
2. **Temporal Distribution**: Years (2018-2025) and quarters assigned randomly
3. **Demographic Generation**: Age → Age group → Sex → Pregnancy status (for applicable females)
4. **Key Population Identification**: Based on NACA estimates (~5% KP prevalence)
5. **HIV Status Determination**: Adjusted for key population status (higher prevalence among KPs)
6. **Treatment Cascade Progression**:
   - Diagnosis → Linkage to care → ART initiation → Time on ART
   - Regimen assignment based on time on ART (increased likelihood of later-line regimens over time)
7. **Adherence and Viral Load**:
   - ART adherence modeled based on regimen type and time on treatment
   - Viral load results generated with 95% suppression rate among those on effective treatment
   - Viral load distribution follows log-normal pattern typical of HIV measurements
8. **Drug Resistance Testing**:
   - Indicated by detectable viral load, treatment failure, or random sampling (15%)
   - DRM probability adjusted for key population, zone, treatment duration, and adherence
   - Specific DRMs selected from common Nigerian mutations
   - Resistance categories assigned based on mutation patterns
9. **Treatment Outcomes**: Determined by viral suppression status and other clinical factors
10. **Health System & Socioeconomic Factors**: Facility characteristics, stockouts, clinical indicators, prevention services, and socioeconomic determinants assigned based on empirical distributions

### Test Standards
- Viral load suppression threshold: 1,000 copies/mL (WHO standard)
- Drug resistance testing: Genotyping method (standard in Nigerian surveillance)
- CD4 counting: Flow cytometry (cells/μL)
- WHO clinical staging: Based on clinical indicators per WHO guidelines

## Dataset Description

The dataset consists of three CSV files representing different burden scenarios:
- `hiv_treatment_cascade_low_burden.csv`: 4,000 records
- `hiv_treatment_cascade_moderate.csv`: 5,000 records  
- `hiv_treatment_cascade_high.csv`: 6,000 records

### Variables and Descriptions

| Variable Name | Type | Description | Range/Values |
|---------------|------|-------------|--------------|
| record_id | String | Unique identifier for each record | Format: NGR-HIV-[SCENARIO]-xxxxxx |
| state | String | Nigerian state (36 states + FCT) | Abia, Adamawa, ..., Zamfara, FCT Abuja |
| lga | String | Local Government Area | State-specific LGA names |
| year | Integer | Year of observation | 2018-2025 |
| quarter | Integer | Quarter of year | 1, 2, 3, 4 |
| zone | String | Geopolitical zone | North Central, North East, North West, South East, South South, South West |
| person_id | String | Anonymized person identifier | Format: PID-xxxxxxx |
| age_years | Integer | Age in years | 15-80 |
| age_group | String | Age category | 15-24, 25-34, 35-49, 50+ |
| sex | String | Biological sex | male, female |
| pregnancy_status | String | Pregnancy status (females 15-49) | yes, no, not_applicable |
| key_population | Boolean | Key population status | True, False |
| kp_type | String | Key population type | fsfw, msm, pid, tg, prisoners (or NULL) |
| hiv_status | String | HIV test result | positive, negative |
| date_of_diagnosis | Date | Date of HIV diagnosis | YYYY-MM-DD format |
| days_since_diagnosis | Integer | Days since HIV diagnosis | 0-3650 (approx. 0-10 years) |
| ever_linked_to_care | Boolean | Ever linked to HIV care | True, False |
| date_of_art_initiation | Date | Date of ART initiation | YYYY-MM-DD format |
| months_on_art | Integer | Months on antiretroviral therapy | 0+ |
| current_art_regimen | String | Current ART regimen | first_line, second_line, third_line |
| art_adherence_pct | Float | ART adherence percentage | 0.0-1.0 |
| missed_doses_last_month | Integer | Number of missed ARV doses in last month | 0+ |
| vl_test_done | Boolean | Viral load test performed | True, False |
| vl_result | Float | Viral load result (copies/mL) | Detectable range (log-normal distribution) |
| vl_suppressed | Boolean | Viral load suppression (<1,000 copies/mL) | True, False, NULL |
| vl_suppression_threshold | Integer | Viral load suppression threshold | 1,000 copies/mL (fixed) |
| vl_detectable | Boolean | Low-level viral detectability | True, False, NULL |
| dr_test_done | Boolean | Drug resistance test performed | True, False |
| dr_test_method | String | Drug resistance testing method | genotyping (or NULL) |
| num_drms_detected | Integer | Number of DRMs detected | 0-4 |
| high_level_risk_drugs | String | Drugs with high-level resistance | Comma-separated drug names (or empty/NULL) |
| specific_drms | String | Specific DRMs detected | Comma-separated mutation codes (or empty/NULL) |
| dr_category | String | Drug resistance category | none, low, intermediate, high |
| transmitted_resistance_flag | Boolean | Likelihood of transmitted resistance | True, False, NULL |
| treatment_outcome | String | HIV treatment outcome | suppressed, failed, defaulted, died, lost_to_followup |
| facility_level | String | Health facility level | primary, secondary, tertiary |
| managing_authority | String | Facility managing authority | public, private, ngo, faith_based |
| urban_rural | String | Geographic setting | urban, rural |
| stockout_arv | Boolean | ARV stockout at facility | True, False |
| stockout_rt_kits | Boolean | Rapid test kit stockout at facility | True, False |
| cd4_count_cells_ul | Integer | CD4 count (cells/μL) | 0+ (or NULL) |
| who_clinical_stage | Integer | WHO clinical HIV stage | 1, 2, 3, 4 (or NULL) |
| tb_screening_done | Boolean | TB screening performed | True, False |
| tb_preventive_therapy | Boolean | TB preventive therapy received | True, False |
| cervical_cancer_screening | Boolean | Cervical cancer screening (females ≥30) | True, False |
| mental_health_screening | Boolean | Mental health screening performed | True, False |
| household_wealth_quintile | String | Household wealth quintile | poorest, poorer, middle, richer, richest |
| education_level | String | Highest education level | none, primary, secondary, tertiary |
| distance_to_facility_km | Float | Distance to health facility (km) | 0.0-50.0 |
| transport_cost_ngn | Float | Transport cost to facility (Nigerian Naira) | 0.0+ |
| stigma_experienced | Boolean | Experienced HIV-related stigma | True, False |
| disclosure_status | String | HIV status disclosure level | full, partial, none, not_applicable |
| partner_notification_done | Boolean | Partner notification completed | True, False |
| condom_use_at_last_sex | Boolean | Condom use at last sexual encounter | True, False |

### Sample Data (First 5 rows from high burden scenario)
```
record_id,state,lga,year,quarter,zone,person_id,age_years,age_group,sex,pregnancy_status,key_population,kp_type,hiv_status,date_of_diagnosis,days_since_diagnosis,ever_linked_to_care,date_of_art_initiation,months_on_art,current_art_regimen,art_adherence_pct,missed_doses_last_month,vl_test_done,vl_result,vl_suppressed,vl_suppression_threshold,vl_detectable,dr_test_done,dr_test_method,num_drms_detected,high_level_risk_drugs,specific_drms,dr_category,transmitted_resistance_flag,treatment_outcome,facility_level,managing_authority,urban_rural,stockout_arv,stockout_rt_kits,cd4_count_cells_ul,who_clinical_stage,tb_screening_done,tb_preventive_therapy,cervical_cancer_screening,mental_health_screening,household_wealth_quintile,education_level,distance_to_facility_km,transport_cost_ngn,stigma_experienced,disclosure_status,partner_notification_done,condom_use_at_last_sex
NGR-HIV-HIGH-000001,Cross River,Cross River_LGA_4,2020,3,South South,PID-6224156,36,35-49,male,not_applicable,False,,negative,,,,,,,,,,,,1000,,,,,,,,,,primary,public,rural,False,False,,,True,False,False,True,richer,primary,18.75,1593.07,False,not_applicable,False,False
NGR-HIV-HIGH-000002,Kaduna,Kaduna_LGA_9,2020,1,North West,PID-3707940,49,35-49,female,no,False,,negative,,,,,,,,,,,,1000,,,,,,,,,,primary,ngo,urban,False,False,,,True,False,False,False,richer,secondary,0.06,8.89,False,not_applicable,False,True
NGR-HIV-HIGH-000003,Plateau,Plateau_LGA_5,2024,2,North Central,PID-8210935,17,15-24,female,no,False,,negative,,,,,,,,,,,,1000,,,,,,,,,,primary,public,urban,False,False,,,False,False,False,True,richer,none,9.6,549.59,False,not_applicable,False,True
NGR-HIV-HIGH-000004,Kwara,Kwara_LGA_8,2020,1,North Central,PID-6783656,29,25-34,male,not_applicable,False,,negative,,,,,,,,,,,,1000,,,,,,,,,,tertiary,public,urban,False,False,,,True,False,False,True,richer,primary,1.92,157.24,True,not_applicable,False,True
NGR-HIV-HIGH-000005,Oyo,Oyo_LGA_6,2021,4,South West,PID-3568714,41,35-49,female,no,False,,negative,,,,,,,,,,,,1000,,,,,,,,,,primary,public,rural,False,False,,,True,False,False,True,middle,secondary,12.06,1025.10,False,not_applicable,False,True
```

## Statistical Distributions

### Viral Load Measurements
Viral load measurements follow a log-normal distribution, which is characteristic of HIV viral load data:
- For suppressed cases: Mean = 3.0 (log scale), Sigma = 0.5 → Median ~20 copies/mL
- For unsuppressed cases: Mean = 4.5 (log scale), Sigma = 1.5 → Wider distribution with higher values
- Values are back-transformed to copies/mL for reporting

### ART Adherence
ART adherence follows a truncated normal distribution:
- First-line regimens: Mean = 0.82, Sigma = 0.15, truncated to [0,1]
- Second-line regimens: Mean = 0.75, Sigma = 0.15, truncated to [0,1]
- Third-line regimens: Mean = 0.70, Sigma = 0.15, truncated to [0,1]

### CD4 Count
CD4 count follows a log-normal distribution:
- Untreated/early treatment: Mean = 5.0 (log scale), Sigma = 1.0 → Median ~148 cells/μL
- On treatment (>12 months): Mean = 6.0 (log scale), Sigma = 0.8 → Median ~403 cells/μL

### Distance to Facility
Distance to health facility follows an exponential distribution:
- Scale parameter = 5.0 km
- Maximum capped at 50.0 km
- Reflects realistic accessibility patterns where most people live near facilities but some face significant barriers

## Source Attribution

### Primary Sources
1. National Agency for the Control of AIDS (NACA). Nigeria HIV/AIDS Indicator and Impact Survey (NAIIS) 2018. Available at: https://naca.gov.ng
2. National Agency for the Control of AIDS (NACA). Press Conference on HIV/AIDS Response Progress, November 2025. Available at: https://naca.gov.ng/news
3. The Sun Nigeria. "Nigeria records 46% drop in new HIV infections." November 15, 2025. Available at: https://sunnewsonline.com
4. Guardian Nigeria. "95% of Nigerians on HIV treatment can no longer transmit virus." November 20, 2025. Available at: https://guardian.ng
5. Patel et al. (2026). HIV Drug Resistance Among Key Populations in Nigeria: Results from a Multi-site Surveillance Study. Journal of Acquired Immune Deficiency Syndromes. PMID: 40931993. DOI: 10.1097/QAI.0000000000003763
6. World Health Organization. Consolidated guidelines on HIV prevention, testing, treatment, service delivery and monitoring: recommendations for a public health approach. 2021 edition. Geneva: WHO; 2021. Licence: CC BY-NC-SA 3.0 IGO.
7. United Nations Programme on HIV/AIDS (UNAIDS). Global AIDS Update 2025. Available at: https://unaids.org

### Additional References
8. Nigeria Federal Ministry of Health. National Guidelines for HIV Prevention, Treatment and Care. 2020.
9. Centers for Disease Control and Prevention (CDC). HIV Drug Resistance Surveillance. Available at: https://cdc.gov/hiv/drugresistance
10. International AIDS Society. HIV Drug Resistance Report 2025. Available at: https://ias.org

## Usage Examples

### Loading the Dataset in Python
```python
import pandas as pd

# Load one of the scenarios
df_high = pd.read_csv('hiv_treatment_cascade_high.csv')
df_moderate = pd.read_csv('hiv_treatment_cascade_moderate.csv')
df_low = pd.read_csv('hiv_treatment_cascade_low_burden.csv')

# Combine all scenarios for comprehensive analysis
df_all = pd.concat([df_high, df_moderate, df_low], ignore_index=True)

# Basic dataset information
print(f"Dataset shape: {df_all.shape}")
print(f"Columns: {df_all.columns.tolist()}")
print(f"HIV prevalence: {df_all['hiv_status'].value_counts(normalize=True)['positive']:.3f}")
```

### Analyzing Treatment Cascade
```python
# Calculate 95-95-95 metrics
total_population = len(df_all)
know_status = df_all[df_all['hiv_status'] == 'positive'].shape[0]
on_treatment = df_all[df_all['ever_linked_to_care'] == True].shape[0]
virally_suppressed = df_all[(df_all['vl_test_done'] == True) & 
                           (df_all['vl_suppressed'] == True)].shape[0]

# Among those who know status
know_status_pct = know_status / total_population * 100
on_treatment_pct = on_treatment / know_status * 100 if know_status > 0 else 0
virally_suppressed_pct = virally_suppressed / on_treatment * 100 if on_treatment > 0 else 0

print(f"Know status: {know_status_pct:.1f}%")
print(f"On treatment: {on_treatment_pct:.1f}%")
print(f"Virally suppressed: {virally_suppressed_pct:.1f}%")
```

### Analyzing HIV Drug Resistance
```python
# Focus on those who received drug resistance testing
dr_tested = df_all[df_all['dr_test_done'] == True]
dr_prevalence = dr_tested[dr_tested['num_drms_detected'] > 0].shape[0] / dr_tested.shape[0] * 100

print(f"Overall DRM prevalence among tested: {dr_prevalence:.1f}%")

# DRM prevalence by key population
kp_dr_tested = dr_tested[dr_tested['key_population'] == True]
kp_dr_prevalence = kp_dr_tested[kp_dr_tested['num_drms_detected'] > 0].shape[0] / kp_dr_tested.shape[0] * 100 if kp_dr_tested.shape[0] > 0 else 0

print(f"DRM prevalence among key populations: {kp_dr_prevalence:.1f}%")

# DRM prevalence by zone
zone_dr = dr_tested.groupby('zone').apply(
    lambda x: x[x['num_drms_detected'] > 0].shape[0] / x.shape[0] * 100 if x.shape[0] > 0 else 0
)
print("DRM prevalence by zone:")
print(zone_dr.round(1))

# Most common DRMs
all_drms = []
for drms_str in dr_tested['specific_drms'].dropna():
    if drms_str:  # Not empty
        all_drms.extend(drms_str.split(','))
        
from collections import Counter
drms_counter = Counter(all_drms)
print("Most common DRMs:")
for drm, count in drms_counter.most_common(5):
    print(f"  {drm}: {count}")
```

### Visualizing Treatment Outcomes
```python
import matplotlib.pyplot as plt
import seaborn as sns

# Treatment outcomes by regimen
plt.figure(figsize=(10, 6))
outcome_regimen = pd.crosstab(df_all[df_all['hiv_status'] == 'positive']['current_art_regimen'], 
                              df_all[df_all['hiv_status'] == 'positive']['treatment_outcome'],
                              normalize='index') * 100
outcome_regimen.plot(kind='bar', stacked=True)
plt.title('Treatment Outcomes by ART Regimen')
plt.ylabel('Percentage')
plt.xlabel('ART Regimen')
plt.legend(title='Treatment Outcome', bbox_to_anchor=(1.05, 1), loc='upper left')
plt.tight_layout()
plt.show()

# Viral suppression over time
plt.figure(figsize=(10, 6))
suppression_by_year = df_all[(df_all['hiv_status'] == 'positive') & 
                            (df_all['vl_test_done'] == True)].groupby('year')['vl_suppressed'].mean() * 100
plt.plot(suppression_by_year.index, suppression_by_year.values, marker='o')
plt.title('Viral Suppression Trends Over Time')
plt.ylabel('Percentage Virally Suppressed')
plt.xlabel('Year')
plt.grid(True, alpha=0.3)
plt.show()
```

## Limitations and Caveats

1. **Synthetic Nature**: While based on empirical evidence, this dataset is synthetically generated and may not capture all nuances of real-world HIV program data.

2. **Parameter Uncertainty**: Some parameters (particularly those related to drug resistance) have confidence intervals in the source studies that are not fully propagated in the synthetic data generation.

3. **Temporal Dynamics**: The dataset assumes stable parameter values over the 2018-2025 period, whereas in reality, HIV epidemiology and program performance may change over time.

4. **Geographic Representation**: While all Nigerian states are included, the distribution may not perfectly reflect actual population sizes or HIV burden distribution.

5. **Key Population Definition**: Key population identification is based on self-reported behavior in the source studies; actual key population sizes may vary by location and methodology.

6. **Drug Resistance Testing**: In real programs, drug resistance testing is not universally available; the simulated testing rates may overestimate actual coverage in some settings.

7. **Behavioral Factors**: Sexual behavior, substance use, and other risk factors are simplified in the dataset and may not capture complex real-world patterns.

8. **Health System Factors**: Stockout rates and facility characteristics are based on available reports but may vary significantly between facilities and over time.

9. **Missing Data Patterns**: The dataset uses specific missing data mechanisms that may not fully reflect real-world data collection challenges.

10. **Causal Interpretations**: While the dataset incorporates logical relationships, users should be cautious about making causal inferences without appropriate study designs and controls.

## License

This dataset is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). You are free to:

- **Share** — copy and redistribute the material in any medium or format
- **Adapt** — remix, transform, and build upon the material for any purpose, even commercially

Under the following terms:
- **Attribution** — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the endorser endorses you or your use.

To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

## Citation

If you use this dataset in your research, please cite it as follows:

```bibtex
@dataset{nigeria_hiv_treatment_cascade_2026,
  author = {HIV Data Synthesis Team},
  title = {HIV Treatment Cascade and Drug Resistance Dataset for Nigeria},
  year = {2026},
  publisher = {HIV Research Consortium},
  version = {1.0},
  url = {https://github.com/your-repo/nigeria-hiv-datasets},
  doi = {10.5281/zenodo.XXXXXXX},
  note = {Synthetic dataset generated based on evidence from NACA reports, UNAIDS 95-95-95 progress monitoring, and published literature on HIV drug resistance in Nigeria.}
}
```

### Alternative Citation Formats

**APA:**
HIV Data Synthesis Team. (2026). HIV Treatment Cascade and Drug Resistance Dataset for Nigeria (Version 1.0) [Dataset]. HIV Research Consortium. https://doi.org/10.5281/zenodo.XXXXXXX

**MLA:**
HIV Data Synthesis Team. "HIV Treatment Cascade and Drug Resistance Dataset for Nigeria." HIV Research Consortium, 2026, HIV Research Consortium, doi:10.5281/zenodo.XXXXXXX.

**Chicago:**
HIV Data Synthesis Team. 2026. HIV Treatment Cascade and Drug Resistance Dataset for Nigeria. Version 1.0. HIV Research Consortium. https://doi.org/10.5281/zenodo.XXXXXXX.

---

*Dataset generated on March 14, 2026 using the NigeriaHIVDatasetGenerator class. For questions or feedback regarding this dataset, please contact the HIV Data Synthesis Team.*