Aims To ascertain the frequency and identify predictors of self-reported hypoglycaemia in Type 1 and insulin-treated Type 2 diabetes. Methods A random sample of 267 people with insulin-treated diabetes were recruited from a population-based diabetes register in Tayside, Scotland. Each subject prospectively recorded the number of mild and severe hypoglycaemic episodes experienced over a 1-month period. Ordinal logistic regression was performed to identify potential predictors of hypoglycaemia. Results Five hundred and seventy-two hypoglycaemic events were reported by 155 patients. The participants with Type 1 diabetes had a total of 336 hypoglycaemic events with a rate of 42.89 events per patient per year. Of these, nine were severe hypoglycaemic events, with a rate of 1.15 events per patient per year. Participants with insulin-treated Type 2 diabetes experienced a total of 236 hypoglycaemic events with a rate of 16.37 events per patient per year. Of these, five were severe hypoglycaemic events, which would be equivalent to 0.35 events per patient per year. Predictors of hypoglycaemia in Type 1 diabetes were a history of previous hypoglycaemia ( P = 0.006) and co-prescribing of any oral drug ( P = 0.048). In patients with insulin-treated Type 2 diabetes, a history of previous hypoglycaemia ( P < 0.0001) and duration of insulin treatment ( P = 0.014) were significant predictors. Conclusion The incidence of self-reported severe hypoglycaemia in insulintreated Type 2 diabetes is lower than in Type 1 diabetes but does occur more often than previously reported and with sufficient frequency to cause significant morbidity. Duration of insulin treatment is a key predictor of hypoglycaemia in insulin-treated Type 2 diabetes.