Invasive pneumococcal disease in Victoria: a better measurement of the true incidence

Abstract

Invasive pneumococcal disease (IPD) notifications are used to monitor IPD vaccination programmes. We conducted sequential deterministic data-linkage between IPD notifications and hospitalization data in Victoria, Australia, in order to determine whether all diagnosed cases were being reported. The proportion of each relevant hospital admission ICD-10-AM code that could be linked to notified cases was calculated. Total and age-specific annual rates were calculated and compared for notified and non-notified cases. Total incidence was estimated using data-linkage results and application of a two-source capture-recapture method. The first 2 years of IPD surveillance in Victoria missed at least one-sixth of laboratory-confirmed IPD cases. Estimated annual IPD rate increased from 9.0 to 10.7/100,000 and rose even higher, to 11.5/100,000, with age-specific rates possibly reaching 90.0/100,000 children aged <2 years, when using capture-recapture. Strategies to improve notification and coding of hospitalized cases of IPD are required.