Oral abstracts Inpatient services A1 (P465) Assessment of hyperglycaemia in emergency hospital admissions using HbA1c: a retrospective audit over one year S Ghosh 1,4,8 , PW Manning 2 , PG Nightingale 3,4 , I Alonso-Perez 5 , F Evison 5 , IM Stratton 6,4 , G Roberts 7,4 and S Manley 4,8 1 Diabetes Centre, University Hospitals Birmingham NHS Trust, Birmingham, UK, 2 Clinical Laboratory Services, University Hospitals Birmingham NHS Trust, Birmingham, UK, 3 Wellcome Trust Clinical Research Facility, University Hospitals Birmingham NHS Trust, Birmingham, UK, 4 Diabetes Translational Research Group, Queen Elizabeth Hospital Birmingham, Birmingham, UK, 5 Health Informatics, University Hospitals Birmingham NHS Trust, Birmingham, UK, 6 Gloucester Diabetic Retinopathy Research Group, Gloucestershire Hospitals NHS Foundation Trust, Cheltenham, UK, 7 University Hospital Waterford, Waterford, Ireland, 8 Medical School, University of Birmingham, Birmingham UK Introduction: Although HbA1c is recommended by NICE and American Diabetes Association for screening medical admissions for diabetes, there is still a paucity of evidence for its use for this purpose in hospital patients. Methods: Emergency hospital admissions were audited between April 2014 and March 2015, n = 43,201, with HbA1c requested for routine care in 8% whereas admission plasma glucose (APG) was measured in 74% of patients. Results: When those with glucose measured on admission, 31,927 (74%), were compared to those with HbA1c, 3,409 (8%), differences were observed in age 62 (43–78) vs 66 (51–78) years, respectively, median and inter-quartile range, p < 0.001; gender 50% vs 48% female, p = 0.002; BMI 26 (22–30) vs 27 (23–32) kg/ m 2 ,p < 0.001; and ethnicity White Caucasian 78%/14% South Asian/5% Afro-Caribbean/4% other vs 76/17/5/3%, p < 0.001. HbA1c was 45 (38–61) mmol/mol with 43% ≥48, 2% ≥120 and 0.3% <20 with linear regression equation log 10 (HbA1c) = 0.476 log 10 (RPG) + 1.253, Spearman’s rho = 0.58, p < 0.001. More patients in HbA1c group had ICD-10 diabetes codes 36% previously/15% during admission/49% never vs 17%/7%/75%, p < 0.001. Admission plasma glucose was also higher 7.6 (5.9– 11.2) vs 6.4 (5.4–8.0)mmol/l, p < 0.001; 25% had glucose ≥11.1 vs 10% and 48% ≥7.8 vs 28%, with 43% with glucose ≥11.1 not diagnosed with diabetes compared to 63% when ≥7.8. Conclusions: Routine HbA1c data show that HbA1c could be used for diagnostic purposes in acute medical admissions with identification of conditions that compromise HbA1c requiring further attention. A2 (P374) Implementation of a hospital inpatient integrated electronic blood glucose and diabetes prescription chart with automated pharmacy insulin alerts can lead to significant and sustained reduction in diabetes related prescription errors NJ Leech 1 , RJ Cook 1 and A Heed 2 1 Diabetes, Newcastle upon Tyne NHS Foundation Trust, Newcastle upon Tyne, UK, 2 Pharmacy, Newcastle upon Tyne NHS Foundation Trust, Newcastle upon Tyne, UK Aim: To determine if reduction in diabetes related prescribing errors observed eight weeks after introduction of an electronic inpatient diabetes monitoring and prescription chart are main- tained at 14 months Methods: We used consecutive National Diabetes Inpatient Audit (NADIA) data collected on all inpatients with diabetes on a single day each year, to quantify diabetes related prescription and management errors before and after the introduction of an inpatient electronic blood glucose and prescribing system with integrated pharmacy insulin alerts (EPS). Results: Diabetes related prescription errors decreased from 24/ 151 (15.7%) before EPS to 15/205 (7.3%) eight weeks post introduction of EPS (p = 0.015), with further reduction to 11/234 (4.7%) at 14 months (p = 0.0046). Insulin prescription errors decreased from 15/62 (24%) pre-EPS to 10/71 (14%) post-EPS and further to 7/91 (7.7%) after 14 months (p = 0.017). Insulin errors not eliminated were “not signed as given” (n = 5) “not written up” (n = 3) and “prescribed at the wrong time” (n = 1). Diabetes related management errors did not improve between 2013 and 2015 (30/151, 20% vs 46/205, 22%). Fourteen months post-EPS after introduction of a peri-operative diabetes pathway and increased DSN input, there was a non-significant reduction in diabetes related management errors (46/205, 22% vs 38/234, 16.2%, p = 0.1). Conclusion: Introduction of an integrated electronic blood glu- cose chart and prescribing system resulted in sustained reduction in diabetes related prescribing errors including insulin. Achieving timely adjustment of diabetes medications in inpatients requires further work. ª 2017 The Authors, presented at the Diabetes UK Professional Conference ª 2017 Diabetes UK. Diabetic Medicine, 34 (Suppl. 1), 5–35 5 DIABETICMedicine DOI: 10.1111/dme.13303