OBJECTIVE To document the prevalence of known diabetes and quantity of diabetes care delivered in long-term institutional settings in South Glamorgan, Wales; to measure physical and mental performance of diabetic residents; and to ascertain the level of basic diabetes knowledge of both staff and diabetic residents.

RESEARCH DESIGN AND METHODS A postal questionnaire survey was sent to 31 nursing or dual registered homes and 88 residential homes in South Wales to determine the prevalence of known diabetes. Medical examination and a semistructured interview of 109 diabetic and 106 age- and sex-matched nondiabetic residents was followed by assessment of dependency (Barthel Activities of Daily Living Scale [15]; Behavioral Rating Scale [16]) and cognitive performance (Mini-Mental State Exam [12]).

RESULTS The prevalence of known diabetes was 7.2%. Forty percent of diabetic residents were taking long-acting sulfonylureas. Only half of the residents in both homes had regular blood glucose monitoring. Less than 1 in 10 diabetic residents were being followed in a hospital-based diabetic clinic. Diabetic residents had had more hospital admissions in the preceding year than nondiabetic control subjects, and their length of stay was double that of nondiabetic control subjects (P < 0.05). Health professional input was fragmented, with involvement of community nurses and dietitians being rare. Diabetic residents had significantly higher levels of arterial disease (P < 0.05), foot ulceration (P < 0.01), and dementia (P < 0.01), and more severe cognitive impairment and higher levels of dependency (P < 0.01). Knowledge of diabetes was poor among both diabetic and nondiabetic residents and care staff.

CONCLUSIONS There is an important need for a reappraisal of diabetes care in institutional settings.