OBJECTIVES: Prior studies indicate diabetes-associated microvascular complications drive substantial healthcare cost; however significant gaps in knowledge remain. This study aimed to provide more accurate measurements of the healthcare resource usage associated with diabetes and its microvascular complications, utilising the cross-care-setting ‘Insights for Care’ diabetes dataset. Through the use of clinical and claims data this dataset allows comprehensive cohort definition, while capturing full hospital care costs. METHODS: Insights for Care captures over 93,000 diabetes patients diagnosed since 2008. This retrospective, observational cohort study includes patients with a first diagnosis of nephropathy/neuropathy in 2010-12, with two years’ data record ± this index date. Patients were identified by ICD-10 diagnosis, procedure and pathology readings. Cohort service utilisation, including number of inpatient, outpatient and A&E events per treatment specialty, and secondary / tertiary care costs were calculated for the 2 years post-index date. RESULTS: Despite a local diabetes prevalence of 7%, diabetes patients account for 13% of hospital patients, 18% of activities, 24% of inpatient bed-days and 31% of estimated secondary care spend – significantly more than previous claims-based studies suggest. Within this group, the cohort of 5,781 patients with nephropathy and 828 patients with neuropathy incurred £60 million of costs within the two years immediately post-first diagnosis. Importantly, the economic impact of microvascular disease escalates substantially with complication severity. For example, per patient costs increase 98% upon first onset of moderate nephropathy (without nephrology: £1,842; with moderate nephrology: £3,648), and by 73% following disease progression to severe nephropathy (£6,324). CONCLUSIONS: This study provides a comprehensive measurement of the burden of diabetes on a typical secondary care provider. It also suggests areas where targeted interventions may improve outcomes and lower resource usage. Additionally, the representative diabetes population used ensures findings are relevant to geographies outside the study area.