The literature contains numerous reports of the frequency of severe hypoglycaemia in childhood insulin-dependent diabetes mellitus. Unfortunately, most of these data have been collected in such a manner as to make comparisons between them difficult. Most reports have used small, well-defined populations of children who were participants in larger, complex, clinical studies. Additional difficulties are encountered in the standardization of the definition of severe hypoglycaemia. Some studies have required that the child lose consciousness while others also included children who required assistance with treatment. Even when studies clearly define the representative nature of their patient population and include universally acceptable criteria for identifying a severe hypoglycaemic episode, the frequency of severe hypoglycaemia must be qualified. Severe hypoglycaemia is generated by a series of complex biological, psychological, and behavioural processes. Its frequency must be viewed not as a fixed value, but rather as a dynamic dependent variable.