Cutaneous leishmaniasis is one of the infectious diseases that affects public health and represents a real threat especially in developing countries. The disease is transmitted by the bite of certain species of sandflies and occurs predominantly in warm, humid and tropical climate.
Finding the source of cutaneous leishmaniasis and identifying factors that promote its spread could help to a good prediction of the epidemic in time. The aim of this study is the construction of a statistical model that reproduces the number of affected cases using climate factors influencing the presence of sandflies.
Given the extensive development of the Generalized Linear Models and their performance in modeling count data as well as their adaptation to the problem of overdispersed data, we present the utility and the basic foundations of Poisson and quasi-Poisson regression models. Thereafter, we build a forecasting model that could predict the number of monthly cases of the cutaneous leishmaniasis from climatic factors during the period 2008-2011 in the province of Msila which is one of the Algerian provinces heavily affected by the epidemic in question. In our case the temperature and trend factor were retained in the model. Poisson regression gave a good result after eliminating the effect of overdispersion.