Search. Read. Cite.

Easy to search. Easy to read. Easy to cite with credible sources.

Diabetic Medicine

Year: 2014  |  Volume: 31  |  Issue: 8  |  Page No.: 971 - 975

Use of hospital admissions data to quantify the burden of emergency admissions in people with diabetes mellitus

D. C. Gibbons, M. A. Soljak, C. Millett, J. Valabhji and A. Majeed

Abstract

Aims

Accurate measurement of emergency diabetes admissions is essential for healthcare delivery and research. This study examines whether current approaches to identifying diabetes-related admissions may underestimate the true burden on hospital care.

Methods

Data spanning the period 1 January 2006 to 31 December 2010 inclusive were extracted from Hospital Episode Statistics data for England. Emergency admissions citing diabetes (E10, E11, E13 or E14) in any diagnosis position in adults (≥ 17 years) were included. E10 and E11 were considered analogous to Type 1 and Type 2 diabetes mellitus respectively; E13 and E14 were grouped as ‘other or unspecified’ diabetes mellitus. For admissions citing diabetes multiple times, those with concordant citations were classified as appropriate; discordant citations were assigned to the ‘other or unspecified’ group. Frequencies of diabetes classifications and complications for each diagnosis position and frequencies of all International Classification of Diseases 10th revision codes for the primary diagnosis field were calculated.

Results

In total, 2 443 046 admissions were identified. Diabetes was cited as the primary diagnosis in 6.2% and most commonly cited as the third diagnosis (23.1%). Type 2 diabetes mellitus was the most common (85.0%). The majority of diabetes citations were ‘without complication’ (2 188 965, 89.6%). The most common primary diagnosis was ‘chest pain, unspecified’ (R07.4, 99 678, 4.1%).

Conclusions

Reliance on the primary diagnosis field to identify emergency admissions in people with diabetes grossly underestimates the true burden placed on hospital care and leads to underestimates of effect sizes in studies utilizing admission rates as outcome measures. An alternative strategy to identify such admissions is required.

View Fulltext