Estimated hospitalisations attributable to seasonal and pandemic influenza in Australia: 2001- 2013.

Aye M Moa,Robin M Turner,David J Muscatello,C Raina Macintyre,Dena L Schanzer

doi:10.1371/journal.pone.0230705

Aye M Moa, Robin M Turner + Show 3 more

Open Access

https://doi.org/10.1371/journal.pone.0230705

Copy DOI

Journal: PloS one	Publication Date: Apr 13, 2020
Citations: 9	License type: CC BY 4.0

Affiliation: UNSW Sydney, University of Otago, Arizona State University

Abstract

BackgroundInfluenza continues to cause seasonal epidemics and pandemics in humans. The burden of influenza is underestimated by traditional laboratory-based surveillance, and modelled estimates are required for influenza-attributable morbidity and mortality. We aimed to estimate the influenza-attributable hospitalisation in Australia, by influenza type.MethodsA generalised-additive regression model was used to estimate type- and age-specific influenza-attributable hospitalisation rates per 100,000 population by principal diagnosis in Australia, from 2001 through 2013. Weekly counts of laboratory-confirmed influenza notifications and by type, influenza A and B were used as covariates in the model. Main principal diagnosis categories of interest were influenza and pneumonia and respiratory admissions. A smoothing spline was used to control for unmeasured time varying factors. Results for 2009, in which the pandemic influenza A(H1N1)pdm09 virus circulated, were not included in annual averages and are reported separately.ResultsDuring the study period, the estimated annual average, all-age, annual respiratory hospitalisation rates attributable to seasonal influenza type A, B and total influenza were 45.4 (95% CI: 34.9, 55.9), 32.6 (95% CI: 22.8, 42.4), and 76.9 (95% CI: 73.6, 80.2) per 100,000 population, respectively. During 2009, the estimated total pandemic influenza-attributable, all-age, respiratory hospitalisation rate was 56.1 (95% CI: 47.4, 64.9) per 100,000. Older adults (≥85 years of age) experienced the highest influenza-attributable hospitalisation rates for both seasonal and 2009 pandemic influenza. Collinearity between influenza A and B time series in some years limited the ability of the model to resolve differences in influenza attribution between the two virus types.ConclusionBoth seasonal and pandemic influenza caused considerable morbidity in Australia during the years studied, particularly among older adults. The pandemic hospitalisation rate in 2009 was lower than the average overall annual rate for seasonal influenza, but young to middle aged adults experience a hospitalisation rate similar to that of severe seasonal influenza.

Highlights

Persons of all ages are susceptible to influenza infection [1, 2]
During 2009, the estimated total pandemic influenza-attributable, all-age, respiratory hospitalisation rate was 56.1 per 100,000
The standard errors for calculating the confidence interval (CI) for average hospitalisation rates across studied years were calculated by using the square root of the sum of the squared standard errors of parameter estimates for each annual influenza variable, divided by the number of influenza seasons included in the average, as shown using the formula below

Summary

Introduction

Persons of all ages are susceptible to influenza infection [1, 2]. Influenza is a vaccine preventable disease, and yet the burden of influenza remains high, varying from season to season, with pandemics being an unpredictable but constant risk [3,4,5].Influenza infection is a common illnesses and annual incidence of infection is not captured accurately by routine surveillance methods such as laboratory-confirmed infections, and true estimates of morbidity and mortality attributable to influenza is not known [6]. Statistical modelling has been used to estimate the true burden of disease impact attributable to influenza infection [3, 4, 7, 8]. A respiratory diagnosis such as influenza combined with pneumonia, all respiratory, and non-respiratory such as all cardiovascular and all-cause diagnosis groups are considered to estimate influenza-attributable illnesses for burden of disease studies [1, 3, 9]. A recent global modelling study estimated that approximately 290,000–650,000 respiratory deaths were attributable to seasonal influenza annually [10]. The burden of influenza is underestimated by traditional laboratory-based surveillance, and modelled estimates are required for influenza-attributable morbidity and mortality.

Objectives

Methods

Results

Conclusion