Proactive inpatient diabetes service model of care observed to decrease severe hyperglycaemia and hypoglycaemia in a 'real-world' Australian hospital setting

Background: Hyperglycaemia and hypoglycaemia are associated with adverse outcomes in hospitalised patients. A cluster-randomised trial at an Australian quaternary hospital showed that an early intervention model of inpatient diabetes care decreased hyperglycaemia. This model of care was later implemented as a clinical service at Northern Health, where the inpatient diabetes service (IDS) performed virtual glycaemic surveillance and provided proactive bedside consultations for patients with unstable glycaemia. Aims: To assess glycaemic and clinical outcomes of this IDS model of care in a real-world hospital setting. Methods: We compared hospital-wide cross-sectional cohorts of all inpatients with diabetes identified on two census days: in November 2017 (pre-IDS implementation) and in March 2021 (with IDS care in place). Bedside diabetes consultations, capillary glucose measurements and glycaemic treatment plans were compared. Standard glucometric analyses were performed on capillary glucose measurements to assess the incidence of hyperglycaemia and hypoglycaemia. Results: The proactive IDS model significantly increased bedside consultations by the diabetes team (6% pre-IDS vs 47% with IDS care) and increased inpatient insulin prescribing. Compared to the pre-IDS group (n = 126), the IDS group (n = 139) had a lower incidence of severe hyperglycaemia (patient-days with mean glucose >15 mmol/L: 6.5% vs 4.5%, P = 0.03) and level 2 hypoglycaemia (patient-days with glucose <3 mmol/L: 1.1% vs 0.2%, P = 0.01). In patients younger than 75 years with admission HbA1c >7.0%, more patients had diabetes treatment intensification at hospital discharge with IDS care. Conclusions: Following implementation of a proactive IDS model of care, a lower incidence of severe hyperglycaemia and hypoglycaemia was observed.