Machine Learning Predictions for HIV Patient Loss to Follow-Up in Haiti, 2018-2024

Haiti's national electronic medical record data for 115,822 people living with HIV across 167 health facilities from 2018 to 2024. This dataset was used to train machine learning models, including CatBoost, to forecast individual risk of loss to follow-up. The study by Jiaqin Wu assessed model performance using metrics like F2-score and recall and interpreted key predictors with SHAP values.

Use Cases

• Forecasting individual patient risk of loss to follow-up based on demographic, clinical, and institutional predictors.
• Analyzing the influence of prior visit frequency and viral load test patterns on retention outcomes.
• Evaluating the impact of antiretroviral therapy dispensation patterns on patient follow-up.
• Assessing how facility location correlates with loss to follow-up risk in a national healthcare system.

Strengths

• Contains data from 115,822 patient records, providing substantial scale for analysis.
• Covers a 6-year time range from 2018 to 2024, allowing for longitudinal study.
• Includes data from 167 health facilities, offering a national perspective from Haiti.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset file is 15.1 KB, indicating a very limited scope, likely containing summary or model results rather than raw patient data.

Provenance

Source

Haiti’s national electronic medical record database.

Collection Method

Analysis of routine electronic medical record data for people living with HIV receiving antiretroviral therapy.

Time Range

2018 to 2024

Freshness

Last updated 2026-04-28 05:01:16; freshness should be verified.

Geography

Haiti