Diabetes--inpatient utilisation, costs and data validity. Dunedin 1985-9

to describe the impact of diabetes on the public hospital inpatient services in one New Zealand region over the period 1985-9, from one 450 bedded university teaching hospital in a major urban centre in New Zealand serving a population of approximately 125,000 people.

by the analysis of routinely collected hospital discharge data and the application thereto of the resource utilisation system (RUS). Data validation was performed by means of a survey of clinical notes.

admission, bed utilisation rates and associated costs for diabetes rose over the period by factors of 3.2, 3.8 and 2.8 respectively; due primarily to the management of macrovascular disease in the elderly. The rise was most marked in those 75 years and over; admission rates, bed utilisation rates and costs in this age group being respectively 1.9, 4.3 and 2.9 times greater than the 60 to 74 year age group. For admissions with diabetes as principal diagnosis, admissions for glycaemic control showed a significant decrease in lengths of stay over the period. Admissions with diabetic peripheral vascular disease were cumulatively, and per case, the most costly, the mean cost, at $16,000 per case, being 5.1 times greater than the mean cost of all admissions with diabetes as principal diagnosis. In 1989 diabetes accounted for 5% of the cost of all inpatient care provided by the hospital. These results are an underestimate, as 45% of admissions of people where diabetes should have been cited as a subsidiary diagnosis were omitted from the discharge data.

diabetes is a major consumer of hospital resources in this region. As the majority of diabetes associated resource use is in the elderly, a population projected to increase significantly, consideration needs to be given to the options for the provision of care. More work also needs to be done on the reasons underlying these changes. If routinely collected data is to be used for these purposes then audit of the data must be performed or use should be restricted to certain defined groups, eg, diabetes as principal diagnosis only.