the HADS and PAID questionnaires. Data from the questionnaires were inputted into a database for analysis. Results: Clinically significant anxiety (HADSA) was prevalent in 4% of patients with type 1 and 5.9% of patients with type 2 diabetes. Clinically significant depression (HADSD) was prevalent in 2% of patients with type 1 and 2.4% of patients with type 2 diabetes. High levels (score ≥40) of diabetesdistress were present in 31% (type 1) and 18.3% (type 2) of patients, respectively. Discussion: Clinically relevant levels of diabetesdistress were present at a disproportionately higher prevalence in patients with type 1 and type 2 diabetes in comparison to levels of anxiety and depression. Greater attention may be needed towards identifying and managing diabetesspecific as well as general mental health issues in people with diabetes. In our hospital, we have implemented dedicated psychology input for patients with diabetes.

Background: Prescribing errors are prevalent in hospital settings with provision of feedback recommended to support prescribing of doctors. Feedback on prescribing has been described as feasible and valued but limited by doctors, with pharmacists described as credible facilitators of prescribing feedback. Evidence supporting prescribing feedback has been limited to date. A formalised programme of pharmacist-led prescribing error feedback was designed and implemented to support prescribers. Objective: To evaluate the impact of a prescribing feedback intervention on prescribing error rates and frequency of prescribing error severity and type. Method: Prospective prescribing audits were undertaken across sixteen hospital wards in a UK teaching hospital over a five day period with 36 prescribers in the intervention group and 41 in the control group. The intervention group received pharmacist-led, individualised constructive feedback on their prescribing, whilst the control group continued with existing practice. Prescribing was re-audited after three months. Prescribing errors were classified by type and severity and data were analysed using relevant statistical tests. Results: A total of 5191 prescribed medications were audited at baseline and 5122 post-intervention. There was a mean prescribing error rate of 25.0% (SD 16.8, 95% CI 19.3 to 30.7) at baseline and 6.7% (SD 9.0, 95% CI 3.7 to 9.8) post-intervention for the intervention group, and 19.7% (SD 14.5, 95% CI 15.2 to 24.3) at baseline and 25.1% (SD 17.0, 95% CI 19.8 to 30.6) post-intervention for the control group with a significant overall change in prescribing error rates between groups of 23.7% (SD 3.5, 95% CI, -30.6 to -16.8), t(75) = -6.9, p < 0.05. The frequency of each error type and severity rating was reduced in the intervention group, whilst the error frequency of some error types and severity increased in the control group. Conclusion: Pharmacist-led prescribing feedback has the potential to reduce prescribing errors and improve prescribing outcomes and patient safety.

Aims: To determine the proportion of UK patients with type 2 diabetes who meet the cardiovascular (CV) or combined CV/core eligibility criteria of cardiovascular outcome trials (CVOTs) showing CV benefit (REWIND, LEADER and SUSTAIN6). Methods: Data from adult patients with type 2 diabetes on/ before Jun 2018 were identified from the Clinical Practice Research Datalink GOLD primary care database and linked to Hospital Episode Statistics data (Protocol 19_262). Patient CV and clinical profiles (age, glycosylated haemoglobin [HbA1c], kidney function, body mass index, medication use) were evaluated against the CVOTs’ eligibility criteria. Data were analysed descriptively. Results: The 33,118 patients included in the analysis had a mean (SD) age of 66.0 (13.3) years and were 56.6% male. Two thirds (64.5%) of the patients met the CV criteria for REWIND, versus 43.0% for both LEADER and SUSTAIN6. The proportion of patients satisfying both CV and core criteria was 44.4% for REWIND, 13.3% for LEADER and 13.5% for SUSTAIN6. The proportion who met the CVOTs’ criteria of ‘established CV disease’ and ‘risk factors only’ for REWIND were 22.4% and 42.1%, respectively, versus 38.7% and 4.3% for both LEADER and SUSTAIN6. Mean (SD) HbA1c of the patients was 7.3% (1.5%), equivalent to REWIND (7.3% (1.1%)) but lower than both LEADER and SUSTAIN6 (8.7% (1.5%)). Study findings remained consistent when restricted to GLP1 RA users. Conclusions: Of the CVOTs reviewed, the REWIND criteria captured a greater proportion of the realworld type 2 diabetes population in the UK. This increased representativeness was largely driven by the broader CV eligibility criteria for REWIND. P43 | Rising rates and widening socioeconomic disparities in diabetic ketoacidosis in type 1 diabetes: A nationwide prospective cohort study