<![CDATA[Introduction: In endemic nations like Indonesia, dengue fever remains a serious public health concern. The majority of predictive models ignore modifiable risk factors at the home level in favor of macroenvironmental variables (rainfall and climate). The study aimed to develop a logistic regression model to forecast the presence of Aedes aegypti larvae based on water quality metrics and the household water containers characteristics. Methods: In Banjar Regency, Indonesia, 400 randomly chosen households participated in a cross-sectional survey. Water parameters (pH, temperature, salinity, and dissolved oxygen) and container attributes (color, cover availability, and type of water source) were evaluated. Significant predictors were identified using stepwise logistic regression. Model performance was assessed using the Receiver Operating Characteristic (ROC) curve and the Hosmer–Lemeshow goodness-of-fit test. Results and Discussion: The final model identified three significant predictors: container color (OR=14.45; 95% CI:2.93–71.16; p=0.001), cover availability (OR=8.02; 95% CI:1.53–42.01; p=0.014), and water source type (OR=16.78; 95% CI:3.18–88.44; p=0.001). The model equation was: logit(p)=-4.676+2.820(water source)+2.671(colour) +2.082(cover availability). The model exhibited outstanding discrimination (AUC=0.945; 95% CI:0.899–0.992) and good calibration (Hosmer-Lemeshow p=0.649). Conclusion: This household-based logistic model effectively identifies high-risk larval habitats using simple, context-specific indicators. Container color, cover availability, and water source type were key predictors of Aedes aegypti larvae presence, offering practical value for community vector control and early warning systems in resource-limited settings.]]>
Copyrights © 2025
