Influenza disease burden in Australia

Transcription

1 Influenza disease burden in Australia Influenza Specialist Group Annual Scientific Meeting Feb 5-6, 217 Dr Jean Li-Kim-Moy

2 Acknowledgements Professor Allen Cheng Dr Kevin Yin Cyra Patel: Research officer NCIRS Members of ATAGI & Influenza Working Party Data provision: ABS; AIHW; FluCAN; NNDSS 2

3 Outline: Estimates of national flu burden Sources: National administrative surveillance data - Influenza notifications - Incidence of influenza hospitalisations - Influenza deaths Indigenous vs Non-indigenous burden Sentinel site surveillance data - Hospital-based: FluCAN - GP (ASPREN; SPN(WA); VicSPIN, BEACH) - ED (NSW, NT, Qld, SA, WA) Not outlined in this talk 3

4 Background Influenza is number one notifiable vaccine preventable disease Attack rates estimated at 3.5% of adults each year but up to 15.2% of children % of infected adults develop clinical disease 2 1 Jayasundara et al. BMC Infect Dis. 214;14:67. 2 Carrat et al. American Journal of Epidemiology 28;167:

5 Burden of influenza in Australia remains high, despite targeted vaccination program >65y Medical conditions with increased risk of flu Indigenous 6m-<5y, 15 y Robust burden data needed to inform evaluations of existing & potential new vaccination strategies 5

6 Methods: Estimating flu burden from administrative data Latest surveillance data were used Calculated rates by age groups (stratified based on age indication of vaccine or potential program options) Trends over time including pre-pandemic, pandemic and post-pandemic time periods. Assessed differences in Indigenous and non- Indigenous burden 6

7 Sources of data that inform key parameters Epidemiological measurement Lab-confirmed influenza & influenza-like illness incidence Hospitalisation incidence rate % ICU admission among hospitalised cases Data category Notifications Data from sentinel surveillance systems ICD-coded hospitalisation Laboratory-confirmed influenza hospitalisation Data source & time period NNDSS (excluding 29) GP sentinel surveillance ED sentinel surveillance AIHW NHMD (excluding 29) FluCAN Mortality Years of Life Lost Population mortality rate In-hospital case fatality ratio (CFR) Death certificates ICD-coded hospitalisation ABS Cause of Death data (excluding 29) NHMD (excluding 29) Abbreviation: ABS=Australian Bureau of Statistics, AIHW=Australian Institute of Health & Welfare, NHMD=National Hospital Morbidity Database, ED=emergency department, FluCAN=The Influenza Complications Alert Network, GP=general practice, NNDSS=National Notifiable Diseases Surveillance System 7

8 Influenza notifications Notification rate population infection rate Notifications: cases who seek medical care, with test performed which is positive, leading to notification Influenced by variation in - Health seeking behaviour - Local testing practices - Accessibility to laboratory testing - Preferential testing of high-risk populations - Use of more sensitive tests in recent years 8

9 Number of influenza notifications by year , 1, 1,571 Number of Notifications 8, 6, 4, 59,26 27,213 44,564 28,38 67,74 2, 1,586 9,173 13,469 3,

10 Annual incidence of influenza notification excl. 29 by age 3 2 Rate (per 1,) 1-5m 6-23m 2-4y 5-11y 12-17y 18-64y 65y Average no9fica9on rate Age group 1

11 Annual incidence of influenza notification , by age group & time period 6 5 No2fica2on rate (per 1,) m 6-23m 2-4y 5-11y 12-17y 18-64y 65y

12 Annual incidence of influenza notification excl. 29 in WA & NT*, by Indigenous status No2fica2on rate (per 1,) Indigenous Others Rela9ve rate ra9o (Indigenous vs. Others) Rela2ve rate ra2o (Indigenous vs. others) - 5m 6-23m 2-4y 5-11y 12-17y 18-64y 65y Age group Note: All rate ratios are statistically significant. * Completeness of Indigenous status: 87% in WA; 98% in NT. Other jurisdictions had suboptimal completeness of indigenous status coding. 12

13 Influenza hospitalisation Data: ICD-coded hospitalisation (any diagnosis) Limitations Variations in health service utilisation, admission threshold, diagnostic/coding practice Cannot exclude multiple admission; transfer between hospitals * from Australian Institute of Health & Welfare (AIHW) National Hospital Morbidity Database (NHMD) (both virologically confirmed [J9 or J1] & not confirmed [J11]) 13

14 Number of influenza ICD-coded hospitalisations by year , Numbers of hospitalisations 1, 8, 6, 4, 2, 3,622 3,94 1,868 2,744 1,879 4,384 2,955 7,335 3,18 5,62 9,93 6,

15 Annual incidence of ICD-coded hospitalisation for influenza (any diagnosis) excl. 29, by level of influenza activity 4 No vaccine 3 Vaccine not funded Hospitalisa2on rate (per 1,) 2 Vaccine funded; but severe outcomes 1-5m 6-23m 2-4y 5-11y 12-17y 18-64y 65y High ac9vity years* Moderate ac9vity years* Low ac9vity years* * Level of influenza activity is defined by ATAGI as: 1) Low activity: annual rate <3 times baseline rate; 2) Moderate activity: annual rate 3 to <5 times baseline rate; 3) High activity: 5 times baseline rate. Baseline rate is the off-season rate of hospitalisation during (excluding 29), where offseason refers to the 6 months which have the lowest hospitalisation rates in each year.

16 Annual incidence of ICD-coded hospitalisation for influenza (any diagnosis) , by age group & time period 25 Hospitalisa2on rate (per 1,) 2 15 Increased hospitalisation Sig rates increased post-pandemic vs pre-pandemic in most hospitalisation age groups rate in adults during pandemic 1 5-5m 6-23m 2-4y 5-11y 12-17y 18-64y 65y

17 Annual incidence of ICD-coded hospitalisation for influenza (any diagnosis) , by age group & time period 25 2 Hospitalisa2on rate (per 1,) 15 What role of increased testing? 1 5-5m 6-23m 2-4y 5-11y 12-17y 18-64y 65y

18 Comparison of notification and hospitalisation rates Rate per 1, Notifications Hospitalisation Year 18

19 % lab-confirmed, coded hospitalisations in adults & adolescents progressively increased after pandemic Proportion of ICD-coded hospitalisation for influenza (any diagnosis: J9 11) being virologically confirmed (J9 1), % 8% Propor2on 6% 4% 2% % m 6-23m 2-4y 5-11y 12-17y 18-64y 65y

20 Annual incidence of ICD-coded hospitalisation for influenza (any diagnosis) * by Indigenous status Hospitalisa2on rate (per 1,) Indigenous Others Vaccine funded; Indigenous except <6mo; 5-14 years (RR % 1.3, 1.9) Rate ra9o (Indigenous vs. Others; all sta9s9cally significant) Rela2ve rate ra2o (Indigenous vs. others) - 5m 6-23m 2-4y 5-11y 12-17y 18-64y 65y Age group * AIHW suggests that Indigenous identification data from 21 are acceptable.

21 Influenza deaths Mortality rate attributable to influenza during excl. 29, by data source* 2.5 Statistically significant Death rate (per 1,) Estimates of both sources: highly conservative (UK rates 1x higher; Cromer. J Infec 214) In-hospital death lower as does not incorporate deaths in community..5. 5m 6 23m 2 4y 5 11y 12 17y 18 64y 65y ABS death cer9ficate data AIHW in- hospital death data * Death certificate data from ABS (any cause of death) vs. in-hospital death from AIHW National Hospital Morbidity Database ICD-coded hospitalisation data (any diagnosis) 21

22 Annual mortality rate for influenza (any cause of death, ), by age group & time period Death rate (per 1,) Caution: small numbers Sig increased deaths rate in 65y postpandemic vs prepandemic Sig increased death rate in adults during pandemic 1-5m 6-23m 2-4y 5-11y 12-17y 18-64y 65y In-hospital death from AIHW National Hospital Morbidity Database ICD-coded hospitalisation data (any diagnosis position) 22

23 Annual mortality rate for influenza (any cause of death, ), by age group & year 12 Death rate per 1, Year -5m 6-23m 2-4y 5-11y 12-17y 18-24y 25-49y 5-64y 65-74y 75y All ages 23

24 Annual mortality rate for influenza coded any cause of death, excl. 29, by age group & Indigenous status 12 Caution: small numbers 3 Death rate (per 1,) * Indigenous Others Rate ra9o (Indigenous vs. Others) Vaccine funded; Indigenous except <6mo; 5-14 years 3.4* 3.* Rela2ve rate ra2o (Indigenous vs. others) m 6-23m 2-4y 5-11y 12-17y 18-64y 65y Age group - 3 * Statistically significant (based on binominal exact method). 24

25 In-hospital case-fatality ratio (CFR) among ICD-coded hospitalisation for influenza (any diagnosis) during excl. 29, by age group 4% 3% CFR among ICD-coded hospitalisation: conservative % death due to influenza, due to Not capturing deaths not coded as due to influenza Incl. multiple admissions or hospital transfers 3.5% (42/11568) CFR 2% 1%.2% (6/2767).2% (1/5178).5% (16/357).5% (15/2837).2% (3/1778) 1.% (197/19755) % 5m 6 23m 2 4y 5 11y 12 17y 18 64y 65y 25

26 Data from sentinel surveillance systems FluCAN: Flu Complications Alert Network Network of 17 hospitals across Australia 12% of national bed capacity Adult and paediatric hospitalisations. Standardised definitions Review of all PCR confirmed influenza hospitalisations during influenza season Annual prospective analysis of each season Control group allows VE estimation 26

27 FluCAN 215 season: Highlights 27 influenza-related hospitalisations Overall 2.1% died. Case fatality 3.3% in 65 y.o 7.5% admitted to ICU. Vaccine coverage in hospitalised: 8.2% in 65y.o. 57.9% in non-elderly adults with medical comorbidities 26.9% in children with medical comorbidities Vaccine effectiveness in those targeted of 45% (95%CI: 34-55%) Cheng et al. Commun Dis Intell 216;4(4):E521 E

28 FluCAN : Paediatric highlights Paediatric hospitalisations Confirms high proportion of admissions <5y.o. ~6% prev healthy 8 11% required ICU admission Numbers of admissions Comorbidity No comorbidity Age at admission (years) Children with comorbidities Twice more likely to be admitted to ICU 214: Full or partial vaccine coverage in test-negative 2 12% in all children, 18% in those considered at risk. VE 55.5% 1. Li-Kim-Moy et al. Epidemiology and Infection. 217 In press. 2. Blyth et al. Eurosurveillance. 216;21(3):pii=

29 Limitations Underascertainment Not every patient is tested. - Local testing practices - Availability and ease of laboratory testing - Caution in interpretation of notifications Hospitalisation and deaths data relies on accurate coding - Actual hospitalisation rates up to 2x higher and deaths 4x higher during (Muscatello). 29

30 Linked national datasets Labconfirmed notification (NNDSS) ICD-coded hospitalisation (NHMD from AIHW) Certified death (ABS) ABS=Australian Bureau of Statistics, AIHW=Australian Institute of Health & Welfare NHMD=National Hospital Morbidity Database, NNDSS=National Notifiable Diseases Surveillance System 3

31 Conclusions Burden of season influenza in Australia remains substantial Existing vaccination program targets those aged 65 years & younger individuals with specific risk factors Disproportionate burden (hospitalisations; death) 65 y.o. Young children, especially those <2y Indigenous persons What role? Childhood vaccination Vaccination of all indigenous Despite limitations, these data are valuable for assessing current vaccination strategies and will help to inform future immunisation strategies 31