Table 1_Structural determinants of depressive symptoms among refugees and host communities

South Sudan data from a 2023 Forced Displacement Survey of 3,055 adults, including 2,066 refugees and 989 host community members. The dataset was analyzed by Hyojin Im using machine learning models to examine structural predictors of depressive symptom severity, measured by the PHQ-9. The study was published on figshare in 2026.

Use Cases

• Training predictive models for depressive symptom severity based on social determinants of health.
• Analyzing the relative importance of factors like self-rated health, perceived violence, and poverty on mental health outcomes.
• Comparing mental health predictor profiles between refugee and host community populations.
• Applying SHAP analysis for model interpretability in a public health context.

Strengths

• Nationally representative sample of 3,055 adults from a 2023 survey.
• Includes data from two distinct population groups: 2,066 refugees and 989 host community members.
• Model performance metrics are provided, with XGBoost achieving an R² of 0.247.
• Predictor importance is quantified, with self-rated health status contributing 19.3% of total importance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset is very small at 15.5 KB, suggesting it may contain summary or processed results rather than raw survey data.

Provenance

Source

2023 Forced Displacement Survey in South Sudan.

Collection Method

Nationally representative survey data analyzed with Elastic Net, Random Forests, and XGBoost models.

Time Range

2023

Freshness

Last updated 2026-04-14 04:19:40; freshness should be verified.

Geography

South Sudan