Measuring outcome of diabetes: a retrospective survey.

The feasibility of using hospital discharge data to monitor six outcome indicators for diabetes care was assessed by reviewing retrospectively the occurrence of these events recorded in hospital records. The population studied was that of the Orkney Islands (19,500) over a 10-year period (1976-1985). During that time, 230 Orkney diabetic patients were treated in Orkney and/or Aberdeen hospitals. The six outcome indicators were: hospital admission rates for diabetic patients (609 in 230 patients), hospital admission rates specifically for diabetic ketoacidosis (29 in 20 patients), rates of diabetes-related lower limb amputations (36 in 23 patients), perinatal mortality rates related to maternal diabetes (nil), visual loss or blindness (13 eyes in 9 patients) and end-stage renal failure (one patient). Scottish Morbidity Returns (which collect data similar to that in the Hospital In-patient Enquiry in England) were shown by case note review to underestimate these outcomes by 41%. Cardiovascular disease, cerebrovascular disease, and peripheral vascular disease accounted for half of all in-patient bed use by diabetic patients (6077 of 13,951 days). The routinely available Scottish hospital discharge data, which are collected for a different purpose, are not sufficiently accurate or complete to reflect variations in actual diabetic events.