Embed Size (px)
Citation preview
www.ncirs.usyd.edu.au
Influenza disease burden in Australia
Influenza Specialist Group Annual Scientific Meeting Feb 5-6, 2017 Dr Jean Li-Kim-Moy
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
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)
3
Not outlined in this talk
Background
§ Influenza is number one notifiable vaccine preventable disease
4
1 Jayasundara et al. BMC Infect Dis. 2014;14:670. 2 Carrat et al. American Journal of Epidemiology 2008;167:775-85.
• Attack rates estimated at 3.5% of adults each year but up to 15.2% of children.1
• 55-78% of infected adults develop clinical disease2
§ 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
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
Sources of data that inform key parameters
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
Epidemiological measurement Data category Data source & time
period
Lab-confirmed influenza & influenza-like illness incidence
Notifications NNDSS 2002–2014 (excluding 2009)
Data from sentinel surveillance systems
GP sentinel surveillance
ED sentinel surveillance
Hospitalisation incidence rate ICD-coded hospitalisation AIHW NHMD 2002–2013 (excluding 2009)
% ICU admission among hospitalised cases Laboratory-confirmed influenza hospitalisation FluCAN 2011–2015
Mortality
Years of Life Lost Death certificates ABS Cause of Death data
2006–2013 (excluding 2009) Population mortality rate
In-hospital case fatality ratio (CFR) ICD-coded hospitalisation NHMD
2002–2013 (excluding 2009)
7
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
Number of influenza notifications by year 2006-2015
9
3,320 10,586 9,173
59,026
13,469
27,213
44,564
28,308
67,704
100,571
0
20,000
40,000
60,000
80,000
100,000
120,000
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Num
ber o
f Not
ifica
tions
Annual incidence of influenza notification 2002–2014 excl. 2009 by age
10
0-‐5m 6-‐23m 2-‐4y 5-‐11y 12-‐17y 18-‐64y ≥65y Average no9fica9on rate 241 236 173 114 74 70 89
0
100
200
300
Rate (per 100,000)
Age group
0-‐5m 6-‐23m 2-‐4y 5-‐11y 12-‐17y 18-‐64y ≥65y 2002-‐08 187 149 62 28 24 18 26 2009 354 420 409 491 533 233 71 2010-‐14 306 342 309 229 142 135 161
0
100
200
300
400
500
600
No2fica2on rate (per 100,000)
Annual incidence of influenza notification 2002-2014, by age group & time period
11
Annual incidence of influenza notification 2002–2014 excl. 2009 in WA & NT*, by Indigenous status
12
6.4
3.7
1.4 1.3 1.6
2.9 3.4
0
2
4
6
8
10
0
500
1000
1500
2000
2500
0-‐5m 6-‐23m 2-‐4y 5-‐11y 12-‐17y 18-‐64y ≥65y
Rela2v
e rate ra
2o
(Indigeno
us vs. others)
No2
fica2
on ra
te (p
er 100,000)
Age group
Indigenous Others Rela9ve rate ra9o (Indigenous vs. Others)
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.
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
13
* from Australian Institute of Health & Welfare (AIHW) National Hospital Morbidity Database (NHMD) † (both virologically confirmed [J09 or J10] & not confirmed [J11])
Number of influenza ICD-coded hospitalisations by year 2002-2013
14
3,622 3,904
1,868
2,744
1,879
4,384
2,955
7,335
3,018
5,602
9,930
6,037
0
2,000
4,000
6,000
8,000
10,000
12,000
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Num
bers
of h
ospi
talis
atio
ns
0-‐5m 6-‐23m 2-‐4y 5-‐11y 12-‐17y 18-‐64y ≥65y High ac9vity years* 326 186 62 21 15 29 94 Moderate ac9vity years* 224 128 32 13 11 19 41 Low ac9vity years* 128 74 18 7 5 10 21
0
100
200
300
400
Hospitalisa2on rate (per 100,000)
Annual incidence of ICD-coded hospitalisation for influenza (any diagnosis) 2002–2013 excl. 2009, by level of influenza activity
* 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 2002–2013 (excluding 2009), where ‘off-season’ refers to the 6 months which have the lowest hospitalisation rates in each year.
No vaccine
Vaccine not funded
Vaccine funded; but severe outcomes
Annual incidence of ICD-coded hospitalisation for influenza (any diagnosis) 2002–2013, by age group & time period
0-‐5m 6-‐23m 2-‐4y 5-‐11y 12-‐17y 18-‐64y ≥65y 2002-‐08 165 114 36 10 8 9 21 2009 157 102 41 22 20 30 45 2010-‐13 208 114 43 19 11 20 49
0
50
100
150
200
250
Hospitalisa2on rate (per 100,000)
Sig increased hospitalisation rate in adults during pandemic
Increased hospitalisation rates post-pandemic vs pre-pandemic in most age groups
0-‐5m 6-‐23m 2-‐4y 5-‐11y 12-‐17y 18-‐64y ≥65y 2002-‐08 165 114 36 10 8 9 21 2009 157 102 41 22 20 30 45 2010-‐13 208 114 43 19 11 20 49
0
50
100
150
200
250
Hospitalisa2on rate (per 100,000)
Annual incidence of ICD-coded hospitalisation for influenza (any diagnosis) 2002–2013, by age group & time period
What role of increased testing?
Comparison of notification and hospitalisation rates 2006-2015
18
0
50
100
150
200
250
300
350
400
450
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Rat
e pe
r 100
,000
Year
Notifications
Hospitalisation
Proportion of ICD-coded hospitalisation for influenza (any diagnosis: J9–11) being virologically confirmed (J9–10), 2002–2013
0%
20%
40%
60%
80%
100%
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Prop
or2o
n
0-‐5m 6-‐23m 2-‐4y 5-‐11y 12-‐17y 18-‐64y ≥65y
% lab-confirmed, coded hospitalisations in adults & adolescents progressively increased after pandemic
Annual incidence of ICD-coded hospitalisation for influenza (any diagnosis) 2010–2013* by Indigenous status
* AIHW suggests that Indigenous identification data from 2010 are acceptable.
2.9 2.6
1.6 1.4
1.8
3.5 3.4
0
1
2
3
4
5
0
300
600
900
1200
1500
0-‐5m 6-‐23m 2-‐4y 5-‐11y 12-‐17y 18-‐64y ≥65y
Rela2v
e rate ra
2o
(Indigeno
us vs. others)
Hospita
lisa2
on ra
te (p
er 100
,000
)
Age group
Indigenous
Others
Rate ra9o (Indigenous vs. Others; all sta9s9cally significant)
Vaccine funded; Indigenous except <6mo; 5-14 years (RR 1.6 95% 1.3, 1.9)
Influenza deaths Mortality rate attributable to influenza during 2006-13 excl. 2009, by data source*
21
* 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)
0–5m 6–23m 2–4y 5–11y 12–17y 18–64y ≥65y ABS death cer9ficate data 0.4 0.2 0.2 0.0 0.0 0.2 2.0 AIHW in-‐hospital death data 0.3 0.2 0.2 0.1 0.0 0.2 1.5
0.0
0.5
1.0
1.5
2.0
2.5
Death rate (per 100,000)
Statistically significant
• Estimates of both sources: highly conservative (UK rates 10x higher; Cromer. J Infec 2014)
• In-hospital death lower as does not incorporate deaths in community.
Annual mortality rate for influenza (any cause of death, 2006–2013), by age group & time period
22
0-‐5m 6-‐23m 2-‐4y 5-‐11y 12-‐17y 18-‐64y ≥65y 2006-‐08 0.29 0.31 0.25 0.04 0.02 0.09 1.57 2009 0.00 0.65 0.24 0.10 0.23 0.81 2.40 2010-‐13 0.60 0.16 0.17 0.05 0.04 0.21 2.32
0
1
2
3
4
Death rate (per 100,000)
Caution: small numbers
In-hospital death from AIHW National Hospital Morbidity Database ICD-coded hospitalisation data (any diagnosis position)
Sig increased deaths rate in ≥65y post- pandemic vs pre-pandemic
Sig increased death rate in adults during pandemic
Annual mortality rate for influenza (any cause of death, 2006–2013), by age group & year
23
0
2
4
6
8
10
12
2006 2007 2008 2009 2010 2011 2012 2013
Dea
th ra
te p
er 1
00,0
00
Year
0-5m
6-23m
2-4y
5-11y
12-17y
18-24y
25-49y
50-64y
65-74y
≥75y
All ages
Annual mortality rate for influenza coded any cause of death, 2006–2013 excl. 2009, by age group & Indigenous status
24
* Statistically significant (based on binominal exact method).
5.5* 1.8 0.0 0.0 0.0
3.4* 3.0*
-‐30
-‐20
-‐10
0
10
20
30
0
2
4
6
8
10
12
0-‐5m 6-‐23m 2-‐4y 5-‐11y 12-‐17y 18-‐64y ≥65y
Rela2v
e rate ra
2o
(Indigeno
us vs. others)
Death rate (p
er 100,000)
Age group
Indigenous Others Rate ra9o (Indigenous vs. Others)
Caution: small numbers
Vaccine funded; Indigenous except <6mo; 5-14 years
In-hospital case-fatality ratio (CFR) among ICD-coded hospitalisation for influenza (any diagnosis) during 2002–2013 excl. 2009, by age group
25
0.2% (6/2767)
0.2% (10/5178)
0.5% (16/3507)
0.5% (15/2837)
0.2% (3/1778) 1.0%
(197/19755)
3.5% (402/11568)
0%
1%
2%
3%
4%
0–5m 6–23m 2–4y 5–11y 12–17y 18–64y ≥65y
CFR
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
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
FluCAN 2015 season: Highlights
§ 2070 influenza-related hospitalisations § Overall 2.1% died. Case fatality 3.3% in ≥65
y.o § 7.5% admitted to ICU. § Vaccine coverage in hospitalised:
• 80.2% 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%)
27
Cheng et al. Commun Dis Intell 2016;40(4):E521–E526.
FluCAN 2011-2014: Paediatric highlights
§ Paediatric hospitalisations 2011-20131 • Confirms high proportion of
admissions <5y.o. • ~60% prev healthy • 8–11% required ICU
admission
28
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 0
20
40
60
80
100
120
Age at admission (years)
Num
bers of adm
issions
Comorbidity
No comorbidity
§ Children with comorbidities • Twice more likely to be admitted to ICU
§ 2014: Full or partial vaccine coverage in test-negative2 • 12% in all children, • 18% in those considered at risk. • VE 55.5%
1. Li-Kim-Moy et al. Epidemiology and Infection. 2017 In press. 2. Blyth et al. Eurosurveillance. 2016;21(30):pii=30301.
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 2005-2008 (Muscatello).
29
Linked national datasets
Lab-confirmed
notification (NNDSS)
ICD-coded hospitalisation (NHMD from AIHW)
Certified death (ABS)
30
ABS=Australian Bureau of Statistics, AIHW=Australian Institute of Health & Welfare NHMD=National Hospital Morbidity Database, NNDSS=National Notifiable Diseases Surveillance System
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
§ Despite limitations, these data are valuable for assessing current vaccination strategies and will help to inform future immunisation strategies
31
What role? • Childhood vaccination • Vaccination of all
indigenous
