Diabetes inpatients: a case of lose, lose, lose. Is it time to use a 'diabetes-attributable hospitalization cost' to assess the impact of diabetes?

AIMS: The UK National Health Service in England pays for inpatients using a formula ('tariff'). The appropriateness of the tariff for people with diabetes is unknown. We have compared the tariff paid and costs for inpatients with/without diabetes and tested the concept of a 'diabetes-attributable hospitalization cost'.
METHODS: This was a cross-sectional, retrospective 12-month audit in a single teaching hospital assessing mortality, bed days per annum and 'diabetes-attributable hospitalization cost' (i.e. the proportion of costs for all patients with diabetes in excess of that paid for comparable patients without diabetes).
RESULTS: There were 64 829 inpatient admissions, with 4864 of those coded as having diabetes; 12.9% was estimated to be the number of patients having diabetes but not coded. People with diabetes occupied 13.9% of all bed days and were 18.1% (1.3-37.8%) more likely to die (age adjusted). The mean bed days per annum were greatest among those with (vs. without) diabetes (men 10.9 ± 17.0 vs. 6.3 ± 12.8; women 11.4 ± 19.4 vs. 5.9 ± 11.6; P < 0.001). The greatest excess admission rates were among those aged 25-64 years. The annual mean tariff was greater for those with diabetes (5380 ± 8740) than those without diabetes (3706 ± 6221) (P < 0.001). The overall cost was even higher among those with diabetes: 5835 ± 11 246 vs. 3567 ± 7238 (P < 0.001). The diabetes-attributable hospitalization cost was 46.5% (9 125 085). An HbA(1c) > 10.0% (> 86 mmol/mol) was associated with excess hospitalization.
CONCLUSIONS: Those with diabetes cost more and are more likely to die when inpatients. The tariff paid for diabetes is high, but in this centre less than the actual costs. Approaches known to reduce hospitalization are urgently required.