Assessing disease burden of respiratory disorders in Hong Kong children with hospital discharge data and linked laboratory data
Objectives: To describe the pattern of respiratory disorders in the Hong Kong paediatric population admitted to government hospitals, and to assess the reliability of the diagnoses by linkage with laboratory data. Methods: Discharge diagnoses for all admissions are recorded in a central computerised database, the Clinical Management System. These data were analysed for the inclusive period July 1997 to June 1999. Virology laboratory results from a single hospital were linked to the Clinical Management System diagnostic codes to examine discrepancies in coding specific viral aetiologies. Results: A primary diagnosis of a respiratory disorder was noted in 37.5% (upper respiratory 30.1%, tonsillitis/pharyngitis 10.5%, croup/ laryngitis 2.3%, acute otitis media 2.7%, bronchitis/chest infection 2.6%, bronchiolitis 10.2%, pneumonia 20.9%, influenza 4%, asthma and allergic rhinitis 16.5%), and a primary or secondary diagnosis in 42.5% of children younger than 15 years. The incidence rates of respiratory illness coded as bronchiolitis and influenza were respectively estimated to be 887-979 and 222-381 per 100 000 children under 5 years and 3551-3949 and 415-528 per 100 000 children under the age of 1 year. The percentage of respiratory-associated admissions varied significantly by hospital and detailed analysis of data at one hospital highlighted important discrepancies between discharge diagnosis and laboratory results. Conclusions: These passive surveillance data provide general estimates of the disease burden for respiratory disorders in Hong Kong children. Active surveillance studies are required to provide more accurate estimates of the disease burden. Consideration should be given to enhance the Clinical Management System by routinely linking all laboratory data with discharge diagnosis information, by establishing sentinel surveillance hospitals and by assessing new strategies to standardise coding.