Southern Hemisphere Influenza and Vaccine Effectiveness Research and Surveillance

Huang, Qiu Sue; Turner, Nikki; Baker, Michael G; Williamson, Deborah A; Wong, Conroy; Webby, Richard; Widdowson, Marc-Alain

Firstly, the disease burden of severe influenza is estimated from the hospital surveillance platform. It measured population-based incidence for SARI-associated influenza hospitalizations including ICU admissions and in-hospital deaths as it provided reliable numerators and denominators, thus without a need for additional healthcare utilization surveys. [68] [69] [70] [71] Our first-year findings (30 April 2012 to 28 April 2013) showed that the SARI-associated influenza hospitalization rate was substantial with the overall adjusted annual incidence of 54/ 100 000 persons (manuscript in preparation). This rate was similar to US data on influenza-associated hospitalizations during 1979-2001, with an average annual incidence of 36Á8/ 100 000. 12 The very young (0-4 years) and elderly (≥65 years) had the highest SARI-associated influenza hospitalization rates, consistent with trends identified in international literature, particularly those from developed countries with temperate climates. 10, [72] [73] [74] [75] A high rate of influenza-related hospitalizations and low vaccine uptake (6%) in young children (6 months to 4 years) from SHIVERS led the NZ government to change vaccination policy by extending free influenza vaccination to those in this age group who have been hospitalized or have a history of significant respiratory illness. 76 SARI surveillance is likely to underestimate the true burden of severe influenza resulting in hospitalization. Some patients will present with non-respiratory symptoms or respiratory disease that does not meet the SARI case definition, or stay briefly in emergency department. 63, 66 SHIVERS has begun to address this gap; a pilot study in 2012 testing persons with respiratory disease who did not meet the SARI case definition showed that a small proportion (6%) of non-SARI patients were positive for influenza viruses, compared with 18% of SARI cases (manuscript in preparation). Future work to expand the case definition to all acute hospital admissions in a sample of very young children will further expand our knowledge of the burden of influenza in this important group potentially protected by maternal immunization. Additionally, SARI and associated influenza cases will be linked to the hospital discharge data to determine the accuracy and validity of the discharge data by determining proportions of the principal discharge diagnosis code categories that are SARI and influenza cases. This will help inform modelling studies of ICD-coded data and help provide some validation of these with laboratory-confirmed data. SARI surveillance is also likely to grossly undercount the actual number of influenza-associated deaths because only a minority of influenza-related SARI deaths are correctly diagnosed, tested and recorded as such. Additional influenza deaths resulting from secondary bacterial infections and exacerbation of pre-existing chronic conditions and atypical clinical presentations are not captured. 77 This limitation presents a challenge in accurately measuring influenzarelated mortality. Future work on statistical modelling may allow for indirect estimation of 'excess' mortality attributable to influenza in those broad categories such as pneumonia, respiratory or circulatory deaths during influenza seasons. 78 Secondly, the disease burden of moderate influenza is estimated using data from SHIVERS ILI surveillance. Our findings from the 2013 season (29 April to 29 September) showed that the ILI-associated