Schistosomiasis transmission in Zimbabwe: Modelling based on machine learning.

Area of Science:

• Public Health
• Epidemiology
• Environmental Science

Background:

• Schistosomiasis poses a significant public health challenge in Zimbabwe.
• Environmental and socioeconomic factors influence disease transmission.

Purpose of the Study:

• To investigate schistosomiasis risk prediction in Zimbabwe.
• To identify key factors affecting *Schistosoma haematobium* and *S. mansoni* prevalence.
• To evaluate the effectiveness of mass drug administration (MDA).

Main Methods:

• Literature review of schistosomiasis data (1980-2022).
• Analysis of 26 environmental and socioeconomic variables.
• Application of three machine-learning algorithms (including Random Forest) for prevalence prediction.

Main Results:

• Environmental and socioeconomic factors, including rainfall, elevation, temperature, economic conditions, and MDA implementation history, significantly correlate with schistosomiasis prevalence.
• The Random Forest model accurately predicted prevalence for both *S. haematobium* and *S. mansoni*.
• Current schistosomiasis prevalence is estimated at 19.8% with MDA, and 23.2% without MDA.

Conclusions:

• Mass drug administration alone is insufficient for schistosomiasis elimination in Zimbabwe.
• Comprehensive control measures, including health education, snail control, and environmental management, are essential.
• Schistosomiasis elimination requires integrated strategies beyond current approaches.