Infectious disease transmission
Networks are used to model disease transmission. They consist of nodes
and edges/links. Nodes = individuals, location, groups. Edges/ links = social
tie, relationship, shared resource or location group membership, a shared
parasite.
Networks can reveal complex behaviours/relationships that are not otherwise obvious.
Network structures
1. Random – No underlying structure/behaviour in the system
2. Small world/scale-free – Variation in behaviour. Most nodes have few links, and a few nodes have many links
(BNOC)
3. Ordered – Variation in behaviour but more predictable.
Stanley Milgram’s small world experiment – confirmed everyone was only 6 degrees of separation from anyone in
the world using his social experiment of letter transferring through personal connections only. So why don’t all
diseases become a pandemic? – due to variation in contact rates and susceptibility.
The majority of pathogen transmission occurs in a small proportion of the population. Most people have few
connections and a few people will have many connections – SUPER SPREADERS.
Super-spreaders are highly- connected hosts that exist in every population for every disease.
R0 = mean number of infections caused by an infected individual in a susceptible population.
- >1 = number of new infections grow
- <1 = number of new infections decline
But the R number obscures considerable individual variation - In the case of malaria, the average is 100 infections
but one individual could infect 3000.
Super-spreaders: Super-spreaders influence the
probability of disease emergence in a social network- Knowing where it
starts can help fight the spread
1. Typhoid Mary was the first super-spreader: Asymptomatic +
High contact rate (chef)
2. Esther Mok, SARS super-spreader (2002): Infected 23 people at
the start + High contact rate (Airline attendant)
Identify the super-spreaders control the infection more successfully.
- Mice Hantavirus study: Hantavirus causes pulmonary system with 70% mortality in humans.
o She covered one wild mouse in fluorescent powder to measure connections.
o The most connected and infected were big, old male mice.
o Confirmed the 20:80 rule – 20% of the population accounts for 80% of the infections.
Case study 1 – Ebola 2014: Mysterious disease spreading in Guinea in 2013 turned out to be
Ebola but was not identified till 2014. 11,000 people died in the epidemic.
Patient zero and contact tracing is used to build a network. However, it is very difficult because it relies on people
remembering where they have been - As soon as someone changes behaviour/visits a big event the tracing becomes
extremely difficult.
Case study 2 – Tasmanian devil: Devil facial tumour disease – one of the very few
transmissible cancers (transmitted by a bite). Emerged in 1996, by 2003 the population was in severe decline. Cull
was initiated in 2004, but it did not work because the network was not studied…
When network was studied in 2009 it revealed they had an unusual ‘giant component’ – all the individuals interacted
with each other, resulting in limited potential to control the disease as there are no specific, highly connected age or
sex classes (such as the big, old male mice in hantavirus study).
Case study 3 – Badgers & Bovine TB: Randomised badger culling trial (RBCT). £50 million.
10 years to complete. 3000km2 across the UK.
DID NOT WORK - By disturbing the population with culling, the territories were unclear and so the badgers mixed
more and there was actually a 20% increase in bTB.
Reservoir hosts cannot be treated without knowing how contact patterns might change – the network is dynamic!
, Summary
- Networks are needed to understand complex mixing behaviour.
- Mixing patterns tend to be ‘small world’ – a few individuals have many contacts.
- Super-spreaders will spread disease rapidly.
- Identify super-spreaders to halt epidemics.
- Networks are dynamic and perturbation effects are unpredictable
Networks are used to model disease transmission. They consist of nodes
and edges/links. Nodes = individuals, location, groups. Edges/ links = social
tie, relationship, shared resource or location group membership, a shared
parasite.
Networks can reveal complex behaviours/relationships that are not otherwise obvious.
Network structures
1. Random – No underlying structure/behaviour in the system
2. Small world/scale-free – Variation in behaviour. Most nodes have few links, and a few nodes have many links
(BNOC)
3. Ordered – Variation in behaviour but more predictable.
Stanley Milgram’s small world experiment – confirmed everyone was only 6 degrees of separation from anyone in
the world using his social experiment of letter transferring through personal connections only. So why don’t all
diseases become a pandemic? – due to variation in contact rates and susceptibility.
The majority of pathogen transmission occurs in a small proportion of the population. Most people have few
connections and a few people will have many connections – SUPER SPREADERS.
Super-spreaders are highly- connected hosts that exist in every population for every disease.
R0 = mean number of infections caused by an infected individual in a susceptible population.
- >1 = number of new infections grow
- <1 = number of new infections decline
But the R number obscures considerable individual variation - In the case of malaria, the average is 100 infections
but one individual could infect 3000.
Super-spreaders: Super-spreaders influence the
probability of disease emergence in a social network- Knowing where it
starts can help fight the spread
1. Typhoid Mary was the first super-spreader: Asymptomatic +
High contact rate (chef)
2. Esther Mok, SARS super-spreader (2002): Infected 23 people at
the start + High contact rate (Airline attendant)
Identify the super-spreaders control the infection more successfully.
- Mice Hantavirus study: Hantavirus causes pulmonary system with 70% mortality in humans.
o She covered one wild mouse in fluorescent powder to measure connections.
o The most connected and infected were big, old male mice.
o Confirmed the 20:80 rule – 20% of the population accounts for 80% of the infections.
Case study 1 – Ebola 2014: Mysterious disease spreading in Guinea in 2013 turned out to be
Ebola but was not identified till 2014. 11,000 people died in the epidemic.
Patient zero and contact tracing is used to build a network. However, it is very difficult because it relies on people
remembering where they have been - As soon as someone changes behaviour/visits a big event the tracing becomes
extremely difficult.
Case study 2 – Tasmanian devil: Devil facial tumour disease – one of the very few
transmissible cancers (transmitted by a bite). Emerged in 1996, by 2003 the population was in severe decline. Cull
was initiated in 2004, but it did not work because the network was not studied…
When network was studied in 2009 it revealed they had an unusual ‘giant component’ – all the individuals interacted
with each other, resulting in limited potential to control the disease as there are no specific, highly connected age or
sex classes (such as the big, old male mice in hantavirus study).
Case study 3 – Badgers & Bovine TB: Randomised badger culling trial (RBCT). £50 million.
10 years to complete. 3000km2 across the UK.
DID NOT WORK - By disturbing the population with culling, the territories were unclear and so the badgers mixed
more and there was actually a 20% increase in bTB.
Reservoir hosts cannot be treated without knowing how contact patterns might change – the network is dynamic!
, Summary
- Networks are needed to understand complex mixing behaviour.
- Mixing patterns tend to be ‘small world’ – a few individuals have many contacts.
- Super-spreaders will spread disease rapidly.
- Identify super-spreaders to halt epidemics.
- Networks are dynamic and perturbation effects are unpredictable
