A stochastic epidemic model which allows some infected individuals to
die from the disease is considered. The model includes the general epi
demic and the fatal disease model as particular cases. A set of iterat
ive equations is derived from which the probability distribution of th
e final size of the epidemic can be computed. It is illustrated that t
he fatal disease model is distinct in that it results in nearly all or
hardly anyone being infected in a large population. An easily compute
d estimator for a key parameter of the model is derived and it is illu
strated that it has good efficiency relative to the maximum likelihood
estimator, which is very tedious to compute.