Summary
Objective: To determine factors affecting length of hospitalisation of infants for recurrent
gastroenteritis using linked data records from the Western Australia heath information
system.
Methods: A seven-year retrospective cohort study was undertaken on all infants born in Western
Australia in 1995 who were admitted for gastroenteritis during their first year of
life (n = 519). Linked hospitalisation records were retrieved to derive the outcome
measure and other demographic variables for the cohort. Unlike previous studies that
focused mainly on a single episode of gastroenteritis, the durations of successive
hospitalisations were analysed using a proportional hazards model with correlated
frailty to determine the prognostic factors influencing recurrent gastroenteritis.
Results: Older children experienced a shorter stay with an increased discharge rate of 1.9%
for each month increase in admission age. An additional comorbidity recorded in the
hospital discharge summary slowed the adjusted discharge rate by 46.5%. Aboriginal
infants were readmitted to hospital more frequently, and had an adjusted hazard ratio
of 0.253, implying a much higher risk of prolonged hospitalisation compared to non-Aborigines.
Conclusions: The use of linked hospitalisation records has the advantage of providing access to
hospital-based population information in the context of medical informatics. The analysis
of linked data has enabled the assessment of prognostic factors influencing length
of hospitalisations for recurrent gastroenteritis with high statistical power.
Keywords
Australian Aborigines - correlated frailty - data linkage - recurrent gastroenteritis
