OA Epidemiology

Mathematical Modelling for Onchocerciasis Control in Africa: Model Predictions, Actual Trends, and Remaining Challenges

Proceedings of the 2013 annual meeting of the Netherlands Epidemiology Society

Volume 1 Issue S1 Abstract 8

 

Wilma A. Stolk, Department of Public Health, Erasmus MC, Rotterdam, the Netherlands
Luc E. Coffeng, Department of Public Health, Erasmus MC, Rotterdam, the Netherlands
Sake J. de Vlas, Department of Public Health, Erasmus MC, Rotterdam, the Netherlands
Jan H.F. Remme, Consultant, Ornex, France

Background
Large scale onchocerciasis control programmes have been on-going since 1975. The stochastic, individual-based microsimulation model ONCHOSIM was developed  in 1990 to support planning and evaluation of control operations, and has been used for over two decades now. Recent epidemiological surveys provide a unique opportunity to evaluate the accuracy of ONCHOSIM predictions on the long-term impact of mass ivermectin treatment and elimination prospects.

Methods
We used epidemiological data from the Gambia and Bakoye river basins in Senegal and Mali: these sites have the longest history of 6-monthly or yearly mass ivermectin treatment in Africa (>16 years) and infection has presumably been eliminated.  For 22 villages, we had longitudinal data on the prevalence and intensity of infection. Grouping the villages by pre- control endemicity level and history of control, we simulated the expected trends in
infection indicators over time. Model-predictions were compared to observed trends.

Results
In general, model predictions and observed trends were in good agreement, although sometimes the prevalence and intensity seemed to decline faster in reality than predicted by the model.

Discussion
Deviations between predictions and observations may be caused by between village variation in timing and coverage of mass treatment. Selection bias might explain why the observed decline in infection indicators was sometimes faster than predicted. However, it is also possible that model assumptions need to be refined. Some uncertainty remains in predicted elimination prospects. We discuss the challenges involved in prediction where and when elimination can be achieved.

Published: 06 Jun, 2013

 
Licensee OA Publishing London 2013. Creative Commons Attribution License (CC-BY)