Endocrine Abstracts

Objectives: To identify patient and clinical factors associated with emergency hospital admission with diabetic ketoacidosis (DKA) in children and young people with type 1 diabetes attending specialist NHS outpatient clinics in England and Wales.

Design: Cross-sectional observational analysis of linked data from the National Paediatric Diabetes Audits (NPDAs) in England and Wales, and Hospital Episode Statistics and Patient Episode Database for Wales. We selected 12 potential independent variables from NPDAs including patient, care process and metabolic control factors. We imputed missing data and fitted Poisson Generalized Linear Multilevel Models with the Paediatric Diabetes Unit (PDU) of treatment as the random effects level.

Setting: PDUs in England and Wales.

Participants: 2944 children and young people with type 1 diabetes with at least one DKA hospital admission between 2008 and 2010.

Interventions: Regular diabetes care provided by PDUs.

Main outcome measures: DKA hospital admissions between 2008 and 2010.

Results: In all age groups girls had a higher average number of DKA admissions over the 3 years than boys. Boys and girls aged 15–19 years had the highest DKA admission rates of any age group (342.3 and 627.3 per 1000 respectively). In the multivariate regression model, deprivation (quintile 5 IRR 1.20, 95% CIs 1.09–1.32), Black/Black Caribbean ethnicity (IRR 1.20, 95% CIs 1.02–1.41), female sex (IRR 1.11, 95% CIs 1.04–1.18), haemoglobin A1c (IRR 1.07, 95% CIs 1.05–1.09), and age (incidence rate ratio (IRR) 1.03, 95% CIs 1.01–1.04) were associated with significantly higher DKA admission rates.

Conclusions: Children from Black/Black Caribbean ethnic groups, those residing in deprived areas, young people aged 15–19 years and those diagnosed with diabetes at an earlier age are all at greater risk of being admitted with DKA. The strong association of admission rate with HbA1c level, but not other intermediate outcomes, shows the impact of HbA1c on short-term as well as chronic complications in diabetes. Our analysis points to the need to understand the determinants of the health inequalities identified and to explore interventions to reduce DKA admissions in these groups.