Abstract:Objectives Integrating malaria prediction models into malaria control strategies can help to anticipate the response to seasonal epidemics. This study aimed to explore the possibility of using routine malaria data and satellite-derived climate data to forecast malaria cases in Togo. Methods Generalised additive (mixed) models were developed to forecast the monthly number of malaria cases in 40 health districts and three target groups. Routinely collected malaria data from 2013 to 2016 and meteorological and vegetation data with a time lag of 1 or 2 months were used for model training, while the year 2017 was used for model testing. Two methods for selecting lagged meteorological and environmental variables were compared: a first method based on statistical approach (‘SA’) and a second method based on biological reasoning (‘BR’). Both methods were applied to obtain a model per target group and health district and a mixed model per target group and health region with the health district as a random effect. The predictive skills of the four models were compared for each health district and target group. Results The most selected predictors in the models per district for the ‘SA’ method were the normalised difference vegetation index, minimum temperature and mean temperature. The ‘SA’ method provided the most accurate models for the training period, except for some health districts in children ≥5 years old and adults and in pregnant women. The most accurate models for the testing period varied by health district and target group, provided either by the ‘SA’ method or the ‘BR’ method. Despite the development of models with four different approaches, the number of malaria cases was inaccurately forecasted. Conclusions These models cannot be used as such in malaria control activities in Togo. The use of finer spatial and temporal scales and non-environmental data could improve malaria prediction.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to predict malaria cases in Togo using routine malaria data and satellite - derived climate data. Specifically, the study aims to explore the possibility of using routinely collected malaria data and meteorological and environmental predictors to predict the number of monthly malaria cases in various health districts and target groups in Togo. This helps to respond to seasonal epidemics in advance and reduce the morbidity and mortality of malaria. ### Research Background - **Malaria Epidemic Situation**: Malaria is an infectious disease mainly transmitted by mosquitoes and is widespread in the WHO African Region. In 2020, there were an estimated 228 million malaria cases (95% of global cases). In Togo, despite the large - scale interventions implemented by the National Malaria Control Program (NMCP) over the past 15 years, malaria remains a significant public health and socio - economic burden. - **Seasonal and Geographical Variations**: In Togo, the seasonal and geographical variations of malaria are obvious, especially shortly after the rainy season or during the rainy season, the number of confirmed malaria cases will increase. The seasonal variation in the northern health districts is particularly significant, especially among children under 5 years old. - **Meteorological and Environmental Factors**: Many studies have shown that meteorological and environmental factors (such as temperature, rainfall and vegetation index) are related to malaria cases. These factors are usually considered as significant predictors of malaria seasonality. ### Research Objectives - **Develop Prediction Models**: The study aims to develop malaria prediction models to help malaria control programs respond to seasonal epidemics in advance and reduce the morbidity and mortality of malaria. - **Use of Data**: The study used routine malaria data and meteorological and environmental data from 2013 to 2016, and the data in 2017 was used for model testing. ### Methods - **Data Sources**: - **Malaria Data**: Monthly data on confirmed malaria cases routinely collected by the Togo National Malaria Control Program (NMCP), aggregated by health districts and target groups (children under 5 years old, children aged 5 years and above and adults, pregnant women). - **Meteorological and Vegetation Data**: Meteorological and vegetation data obtained from NASA's POWER project and MODIS satellites, including total precipitation, wet days, relative humidity, air temperature (average, minimum and maximum), wind speed (average, minimum and maximum) and vegetation index (NDVI). - **Statistical Analysis**: - **Model Types**: Generalized additive models (GAM) and generalized additive mixed models (GAMM) were used for prediction, with negative binomial distribution errors. - **Variable Selection Methods**: Two methods for selecting lagged meteorological and environmental variables were compared: a statistical method ('SA') and a method based on biological reasoning ('BR'). - **Model Training and Testing**: Data from 2013 to 2016 were used for model training, and data in 2017 were used for model testing. ### Results - **Prediction Performance**: - **Training Period**: The'statistical method' provided the most accurate model during the training period, with the explained deviation ranging from 41.4% to 81.9%. - **Testing Period**: Although it performed well during the training period, during the testing period, the prediction results were not accurate, especially in some health districts and target groups. - **Prediction Error**: During the 2017 test period, the differences between the predicted number of cases and the actual number of cases varied among different health districts and target groups, with the percentage error ranging from - 33.5% to 112.1%. ### Discussion - **Research Significance**: This is the first study in Togo to attempt to predict malaria cases using meteorological and environmental predictors on a national scale. Although four different methods were used, the prediction results were not satisfactory. - **Improvement Directions**: The study recommends using finer spatial and temporal scales and non - environmental data (such as intervention data or clinical data) to improve the accuracy of malaria prediction. ### Conclusion - **Limitations of the Current Model**: The current model cannot be directly used for malaria control activities in Togo. The model needs to be further improved, especially by considering finer spatial and temporal scales and more non - environmental data.

Exploring malaria prediction models in Togo: a time series forecasting by health district and target group