Transcript Cumulative attack rate(%)

Published in Nature, July 2006. doi: 10.1038
presented by Yuri Yakushko
Individual-based simulation model of
pandemic influenza transmission
Presentation plan
1. Influenza virus
2. Model set-up
3. Efficacies of different policies
Influenza virus
• family ORTHOMYXOVIRIDAE
• genera: INFLUENZAVIRUS A, B and C
• genome: -RNA, 8 segments,
ca. 13,6 kB (Baltimore: V)
Influenza virus
Clancy, S. (2008) Genetics of the influenza
virus. Nature Education 1(1)
www.apaci-flu.com
Influenza virus
Transmission: droplets
Incubation time: 12 h – 3 d
Duration: about 3 days
Can be complicated by
- bacterial superinfection
- myokarditis, encephalitis
High lethality in very young, very old,
very ill
Can cause pandemic
Symptoms
• sudden onset
• high fever with chills
• fatigue
• headache (retrobulbar)
back-, muscle pain
• coughing (dry),
• sore throat
Age dependent mortality in Influenza-pandemie
Lederberg 1997
1918
United States
Registration area
Individual-based simulation model of pandemic
avian influenza transmission
• For US and UK
• Simulation of the spread of a new influenza pandemic via
transmission by:
- household contact
- schools
- working places
- wider community
• Testing of efficacies of different policies (based on the model)
Databases used
• Census 2005 US/UK
total population size, household size, age structure
• National Centre for Educational Statistics 2004
school size, school allocation
• Landscan 2003 (Oakridge National Lab)
model of instantanous population density
• National Household Travel Survey
(nights spent away from home for long distance journeys
by air)
• …
Probability for an individual to commute x km
Example of data match to the simulation model.
a) UK
b) US
Some definitions (baseline assumptions)
1. Case detection: 90% are detected by healthcare system.
2. Clinical attack rate: the proportion of an exposed population at risk
who develop clinical illness during a defined period of time.
3. Peak daily attack rate(%): during an outbreak, the highest proportion
per day of an exposed population at risk who become infected or develop
clinical illness
4. Cumulative attack rate(%): total proportion of population who develop
clinical illness
5. Border controls: reduction in the number of infected individuals entering
the country (in %) from a certain time.
6. Case isolation: reduce the contacts of an infected individual (def. 90%)
starting one day after reporting symptoms and lasting 7 days.
7. Household quarantine: reduce the community contacts of all individuals
in a household containing a clinical case.
8. Vaccine: produced from the pandemic virus and assumed to reduce
susceptibility of those receiving it by 70%, infectivity – by 30%, probability
to become a clinical case – by 50%. Protection is assumed to start 2
weeks after vaccination.
Other assumptions made
• 50% of infected are ill enough to be clinical cases, 90% are detected
• no change in behaviour of healthy
• no cross-immunity to strain in the population
• all ill children stay at home
• 50% of symptomatic adults go to work
• influenza is seeded in the countries via international travel, absence
of global pandemic spread
Basic Reproductive Numer R0
Mean number of new infected individuals created by a newly infected person
in a population of all susceptible people
(so mean number of secondary cases, how “infectious” an individual is)
R0 < 1 – an infection will go down with certaintly.
Values of R0 for some well-known infectious diseases:
• SARS: R0=2-5
• Measles: R0=16-18
• Influenza:
-1918 pandemic R0=1.7-2.0
-1957 pandemic R0=1.5-1.7
Here:
Moderate transmissibility scenario: R0=1.7
High transmissibility scenario:
R0=2.0
Epidemic curves
a US
b UK
Red R0=2.0
Blue R0=1.7
Cumulative and peak daily attack rates
Cumulative attack rate(red): total proportion of
population who develop clinical illness
Peak daily attack rate(blue): during an outbreak,
the highest proportion per day of an exposed
population at risk who become infected or
develop clinical illness
Red initial case
Blue peak of epidemic
Green peak attack rate
R0=2.0
Pandemic dinamics simulation
Grey: population density
Red: area with infective cases
Green: area where pandemic is over
Starting day 0
Social strategies. Impact of border control
HT – high transmissibility scenario
MT – moderate transmissibility scenario
Impact of local travel restrictions
US
Impact of antiviral treatment
GB, US results identical
Black no treatment
Yellow - 30%
Red - 50%
Blue - 70%
Green -90% cases treated
Black no treatment
Red after 2d
Blue after 1d
Green no delay in treatment
after symptoms onset
when 90% cases are treated
Cumulative attack rate(%): total proportion of population
who develop clinical illness
Black
Red
Blue
Green
no isolation
50%
70%
90% cases isolated
Impact of household/socially targeted policies
US
Grey
Red
no intervention
90% pts. treated
proph. of household
after 1d
Blue voluntary household
quarantine
Green red+blue
Grey
Red
Blue
no intervention
reactive school closure
50% of workplaces
closed
Green red+blue + 90% proph.
Peak daily attack rate(%): during an outbreak, the highest proportion
per day of an exposed population at risk who become infected or develop clinical illness
Impact of vaccination
US
Grey
Red
Blue
Green
no vaccination (0-16y)
d30
d60
d90
orange no age priorisation (d60)
purple elderly first (d60)
Vaccination at the rate 1% of the population per day
should start within 2 month of the initial outbreak
(so when the first cases are detected) – not reachable
with current vaccination technologies.
Maximum transmission reduction
if children vaccinated first
Overall impact of different policies taken
Policy 13: 90% case treatment
+ reactive school closure
Policy 19: policy 13 + household proph
Policy 27: policy 19
+ proph of 80% classmates
and colleagues
+ 99% effective border controls
Summary
1. Household quarantine is potentially the most effective social
distance measure, but only if compliance with policy is good
2. Reactive school closure has limited impact on overall attack
rated, but can enhance other policies
3. Only >99% border control can delay the spread
4. Can not be directly used for future pandemics
5. Lack of data from care homes, prisons etc
6. Lack of data of personal protective measures
Thank you for your attention