Assessing efficient patient care: should length of stay be calculated independently of local admission rates?

OBJECTIVE

To compare the length of hospitalisation for infants with bronchiolitis across the Eastern region and to assess the impact of the varying admission rates in each hospital.

DESIGN

Data collection through the Hospital Episode Statistics (HES) using the ICD clinical coding for bronchiolitis across all hospitals in east of England for three winter seasons (October to March for the years 2009/10, 2010/11 and 2011/12).

MAIN OUTCOME MEASURE

Length of hospital stay, corrected to adjust for local population.

RESULTS

Seventeen hospitals across the east of England were included in this study. Overall admission rate (as a percentage of the population) for the region was 3.3% and consistent with national data, but rates within individual hospitals varied between 1.5% and 5.7% over the 3-year period. Bed days per 1000 population ('standardised bed days') per year varied almost fourfold, from 34.5 to 122.3 in different hospitals. Corrected length of stay showed high discordance when compared to average length of stay.

CONCLUSIONS

The average length of stay is substantially affected by admission rates, with hospitals who admit a greater proportion of infants appearing to have a shorter uncorrected length of stay. We propose that a single corrected measure for length of stay should be used when assessing the efficiency of care because it is unaffected by variations in local admission rates and is adjusted for local population size.