Background Asthma exacerbations pose a significant burden to the NHS, however the factors that underpin inpatient length of stay remain poorly understood. We undertook a service improvement project using hospital episode statistics to evaluate the demographics of the asthma hospital exacerbation population in Leicester hospitals and risk factors that influence length of stay.

Methods Hospital episode statistics were generated using an ACCESS database query set capturing a range of clinical meta data including length of stay, ICD-10 coding for two types of asthma admission (asthma and status asthmaticus), sex, age, ethnicity, financial year quarter of admission. In addition comorbidity was calculated by cross referencing spell diagnoses to the Charlson comorbidity index score. This database was subsequently linked to pathology data reporting blood eosinophils, neutrophils, CRP at admission and the highest ever historical blood eosinophil count. A backwards logistic regression model was used to identify factors that increased length of stay beyond the median for the overall population.

Results We identified 719 exacerbations between 1/4/2013- 31/3/2014. The number of exacerbations ranged from 157–228/quarter with the winter quarters (Q3/4) being associated with the highest exacerbation frequency. Exacerbators had a median (IQR) age of 46 (31–64) years, length of stay of 2 (1–5) days, Charlson score 4[4–4]. The majority of patients were female (72%). The percentage of repeat attender was 12% of the total population with the highest percentage in the winter quarters. Logistic regression models identified that ethnicity (non-white/Caucasian), baseline admission CRP, neutrophil count, age, coding of status asthmaticus and Charlson co morbidity index significantly increased length of stay beyond the median of 2 days. In contrast neither admission or highest ever blood eosinophil count influenced the median length of stay.

Conclusions We have identified a number of factors associated with an increased length of stay in asthma hospital exacerbations in Leicester. Our observations support the notion that age, comorbidity and biochemical features of infection rather than eosinophilic inflammation increase length of stay. Further research is required to examine the mechanisms that underpin asthma admissions in this population and to reduce length of stay.