princekhera/uganda-schistosomiasis-analysis

An R-based analysis toolkit to explore schistosomiasis (Schistosoma mansoni) in Uganda. Includes data summaries, temporal and spatial trends, and predictive modeling to support evidence-based decision-making in public health. # 🧪 Schistosoma mansoni in Uganda: Prevalence, Distribution & Determinants This repository contains an R-based analysis of *Schistosoma mansoni* infections in Uganda. Using WHO-AFRO region survey data, this project summarizes prevalence trends, spatial distribution, and key predictors, supporting evidence-based decision-making in public health. ## 📈 Summary of Findings ### 1. 🗂️ Data Summary - **Total surveys conducted**: 2,205 - **Individuals tested**: 181,037 - **Positive S. mansoni cases**: 38,302 - **Data completeness**: 82.75% ### 2. 📊 Prevalence Over Time - Visualized yearly mean prevalence with 95% confidence intervals. - Added yearly population estimates for school-aged children. - **Welch Two Sample t-test** (pre vs. post-2000): - Mean before 2000: **18.98%** - Mean after 2000: **19.98%** - **p = 0.506** (not statistically significant) 📌 *Conclusion*: No statistically significant change in mean prevalence before vs. after 2000. Mean Prevalence with school aged children.pdf ### 3. 🗺️ Regional Prevalence & Uncertainty | Region | Mean Prevalence (%) | 95% CI | Sample Size | |-------------|---------------------|------------------|-------------| | Northern | 24.1 | (22.0, 26.3) | 619 | | Eastern | 19.4 | (17.9, 20.9) | 991 | | Western | 18.0 | (15.5, 20.4) | 498 | | Central | 16.2 | (14.7, 17.7) | 674 | | NA | 39.5 | (30.5, 48.5) | 60 | | Nord Kivu | 0.0 | (0.0, 0.0) | 2 | - **Northern Region** shows the highest well-supported prevalence. - **Central Region** shows the lowest among major regions. ### 4. 🤖 Determinants of Prevalence (Modeling) Four models were trained using an 80/20 train-test split. **Best model**: `Random Forest` (lowest RMSE). | Model | RMSE | |------------------|--------| | Random Forest | 19.71 | | XGBoost …