Outcome indicators for diabetes services--what do we have and what do we need?

Minimizing the need for hospital admissions for hyperglycaemic coma, hypoglycaemic coma and amputation of the lower limbs in patients with diabetes can be regarded as some of the legitimate objectives of a local diabetes service. Routinely collected data are available to calculate rates for such admissions for health service districts and for regions or their equivalents. East Anglian regional rates for admissions mentioning hyperglycaemic coma fell between 1981 and 1986 while rates for those mentioning hypoglycaemic coma rose. Amputation rates remained steady. Between-district variation for all rates was considerable and certain districts showed consistently high rates from year to year for hyperglycaemic coma with others having consistently high rates for amputations. Lack of standardization of case definition and uncertainty about the validity of routinely collected hospital admission data are the most important drawbacks of this approach. With careful interpretation, however, these data provide a possible source for the measurement of the effectiveness of local diabetes services.