SHORT COMMUNICATION

Horizon Scanning for Emergence of New Viruses: FromConstructing Complex Scenarios to Online GamesP. Gale and A. C. Breed

Animal Health and Veterinary Laboratories Agency, Weybridge, New Haw, Addlestone, Surrey, UK

The aim of this study is to draw attention to the devel-

opment of novel proactive approaches for horizon scan-

ning for emerging viruses in the medium to long term.

By definition, viruses yet to emerge are not in the scien-

tific literature as shown, for example, by the recent emer-

gence of Schmallenberg virus in northern Europe, while

others that are known may re-emerge in unexpected ways,

for example, the transmission of Nipah virus from fruit

bats (Pteropus giganteus) to persons through drinking

fresh date palm sap (Luby et al., 2006). Horizon scanning

techniques can be developed to identify novel routes and

sources, some of which will not yet have been considered.

A major challenge for horizon scanning is that it is

against a background of change. The world is changing

and viruses emerge through adapting to new niches aris-

ing from novel combinations of events.

Central to horizon scanning is prediction of the com-

plex scenarios through which viruses could emerge before

they occur. For example, it seems logical when told after-

wards that there could be a link between the use of a

drug (diclofenac) in cattle and an increased incidence of

rabies in humans (Markandya et al., 2008). The use of

diclofenac decimated vulture populations in India result-

ing in a build-up of cattle carcases that served as a food

source for feral dogs, which increased in abundance

resulting in increased incidence of dog bites in humans.

Talking to scientists in different and diverse fields is one

approach to investigating new combinations of events.

For example, the retreat of glaciers because of climate

change on the island of South Georgia in the South

Atlantic is allowing the spread of rats, previously intro-

duced by whalers, to rat-free parts of the island (Prof.

Alan Rodger, British Antarctic Survey, personal commu-

nication). Rats can be vectors of disease and also may

indirectly affect disease transmission by changing the eco-

systems and affecting population dynamics of other spe-

cies. However, being alerted to new complex

combinations of events is a rare luxury, and for horizon

scanning, we need to develop new tools and approaches

to generate complex scenarios for consideration.

One potential approach involves ‘spidergrams’. This

was explored recently at a workshop funded by the Euro-

pean Science Foundation (ESF; Gale and Jansen, 2010).

The idea is based on ‘cut-up’ techniques where multiple

texts are cut up into smaller portions at random and

rearranged to create new and innovative phrases. The

generation of complex scenarios requires a database of

factors including those known to be linked to the emer-

Keywords:

horizon scanning; emerging viruses; online

games

Correspondence:

P. Gale. Animal Health and Veterinary

Laboratories Agency, Weybridge, New Haw,

Addlestone, Surrey, KT15 3NB, UK.

Tel.: +44 (0)1932 357575;

Fax: +44 (0)1932 359429;

E-mail: paul.gale@ahvla.gsi.gov.uk

Received for publication June 1, 2012

doi:10.1111/j.1865-1682.2012.01356.x

Summary

Horizon scanning techniques can be developed to identify novel routes and

sources for the emergence of viruses in the medium to long term. Central to

horizon scanning is prediction of the complex scenarios through which viruses

could emerge before they occur. One approach involves ‘spidergrams’ in which

complex scenarios are generated by combining factors randomly selected from

different categories of events. Spidergrams provide a framework for how differ-

ent factors could interact, irrespective of the virus, and also enable testing of

combinations not previously considered but which would be ‘tested’ in nature

by a virus. The emergence of viruses through new routes is often related to

changes, for example, in environmental and social factors, and the Internet will

undoubtedly be used to identify long-term trends for consideration. In addi-

tion, online games may provide horizon scanners with suggestions for new

routes and strategies that could be used by emerging viruses.

Transboundary and Emerging Diseases

ª 2012 Crown copyright • Transboundary and Emerging Diseases. 1

gence of viruses. Categories for these factors include: zoo-

logical, farming practice, sustainability, environment and

climate, land use, ecological, socio-economic, human

behaviour, political, technology and medicine. By ran-

domly taking factors from each category and combining

them, many complex scenarios can be generated. Most

will probably be implausible, but somewhere in there,

amongst others of potential interest, is the dichlofenac/

rabies scenario described previously. The approach can be

taken further to include branches and extensions. For

example, a build-up of cattle carcasses may not only

increase the abundance of feral dogs (which could lead to

an increase in ectoparasites that may in turn be vectors

for microbial pathogens), but could also result in more

rubbish building up (because of pressure on municipal

services) and increases in rodent populations. Similarly,

combinations can be built-in. Thus, flooding and higher

temperatures at the same time could increase rodent pop-

ulations and arthropod and snail vectors, for example.

Flooding may not only promote mixing of sheep and cat-

tle (which is a factor for malignant catarrhal fever and

fasciolosis) and increase livestock contacts in general, but

also causes power station failure as happened in the UK

in the summer of 2007. This can lead to failure of water

supplies such that people resort to storing water in open

containers that could serve as breeding sites for certain

urban mosquito species (with associated implications for

mosquito-borne viruses). Similarly, other mosquito spe-

cies, for example, Ochlerotatus caspius in England, could

use the floodplains as breeding sites. The main challenge

of the spidergram approach is eliminating those pairings

of factors that are irrational, rather than unlikely, and

identifying those that interact in some way. At the ESF

workshop (Gale and Jansen, 2010), a spidergram was

developed for the effect of environmental temperature on

the emergence of Nipah virus (Fig. 1). A 2 · 2 table was

used to assess the linkages of factors with respect to Ni-

pah virus (Table 1). The level of containment of the pigs

is an important factor in Nipah virus transmission and is

rationally linked to the other six factors displayed that are

linked to temperature. Spidergrams provide a framework

for how different factors could interact, irrespective of the

virus, and also enable the assessment of combinations not

previously considered but which would be tested in nat-

ure by a virus.

Spidergrams also generate novel discussion points and

new perspectives. An unexpected outcome of the ESF

workshop was the proposed use of spidergrams to

develop questionnaires for case–control studies. Some epi-

demiological observations have no obvious explanations,

for example, the increase in hantavirus incidence in Ger-

many in older persons (Faber et al., 2010). At the work-

shop, a spidergram was developed to identify potential

new reasons and routes of rodent exposure for older per-

sons. The focus was on socio-economic factors and

changes in behaviour. Dementia was also added, which

has caused surprise in some audiences as the increase in

hantavirus incidence was observed in people older than

30 years, which is too young for dementia. However, this

is the age group who may be caring for parents with

dementia in a home that has become rodent-infested. The

question in the case–control study could therefore be,

‘Do you care for an elderly relative with dementia?’ Thus,

links which at first seem inappropriate may prove relevant

on deeper consideration.

Web search data are being used as a proxy to monitor

dengue epidemics (Chan et al., 2011) in the short term.

For the purpose of horizon scanning in the medium to

long term, the Internet is an important resource for iden-

tifying changes and long-term trends that could affect the

emergence of viruses. For example, could the route of

highly pathogenic avian influenza virus, H5N1, through

frozen duck meat to backyard chickens in Germany in

2007 (Harder et al., 2009) have been anticipated through

looking at Internet chat rooms on feeding of meat and

kitchen scraps to backyard poultry? Perhaps of greater

potential application though is using the online commu-

nity to generate spidergrams and ‘design’ new viruses

through online games. The game would include a virus

toolbox and a pathway toolbox. The player would click

on certain icons in the toolbox to build features into the

virus design to allow it to adapt its strategy to the path-

way being designed. Points would be given for the num-

ber of nodes in the spidergram providing the player

could explain each paired linkage. More points would be

given for feedback loops and combinations of interacting

events. Tapping the resource of the online community

has previously been used through a game to determine

the structure of a viral protein (Good and Su, 2011). As

Migration

Level ofcontainment ofpigs/biocontrol

WarHostbehaviour

Poverty

Deforestation

Temperature

Fig. 1. Spidergram generated for potential effect of temperature on

emergence of Nipah virus (adapted from Gale and Jansen, 2010).

Horizon Scanning for Viruses P. Gale and A. C. Breed

2 ª 2012 Crown copyright • Transboundary and Emerging Diseases.

horizon scanners, we would welcome not only the help of

lateral thinkers from extremely diverse backgrounds but

also views on the feasibility of generating spidergrams

through online databases and more importantly spotting

the more meaningful pathways.

References

Chan, E. H., V. Sahai, C. Conrad, and J. S. Brownstein, 2011:

Using web search query data to monitor dengue epidemics:

a new model for neglected tropical disease surveillance. PLoS

Negl. Trop. Dis. 5, e1206.

Faber, M. S., R. G. Ulrich, C. Frank, S. O. Brockmann, G. M.

Pfaff, J. Jacob, D. H. Kruger, and K. Stark, 2010: Steep rise

in notified hantavirus infections in Germany, April 2010.

Euro. Surveill. 15, 19574.

Gale, P., and V. Jansen, 2010: New approaches to horizon

scanning for emerging and infectious viruses in Europe.

Report of ESF exploratory workshop. Available at http://

www.esf.org (accessed June 26, 2012).

Good, B. M., and A. I. Su, 2011: Games with a scientific pur-

pose. Genome Biol. 12, 135.

Harder, T. C., J. Teuffert, E. Starick, J. Gethmann, C. Grund,

S. Fereidouni, M. Durban, K. H. Bogner, A. Neubauer-Juric,

R. Repper, A. Hlinak, A. Engelhardt, A. Nockler, K. Smie-

tanka, Z. Minta, M. Kramer, A. Globig, T. C. Mettenleiter,

F. J. Conraths, and M. Beer, 2009: Highly pathogenic influ-

enza virus (H5N1) in frozen duck carcasses, Germany, 2007.

Emerg. Infect. Dis. 15, 272–279.

Luby, S. P., M. Rahman, J. Hossain, L. S. Blum, M. M. Hus-

ain, E. Gurley, R. Khan, B.-N. Ahmed, S. Rahman, N. Ha-

har, E. Kenah, J. A. Comer, and T. G. Ksiazek, 2006:

Foodbourne transmission of Nipah virus, Bangladesh.

Emerg. Infect. Dis. 12, 1888–1894.

Markandya, A., T. Taylor, A. Longo, M. N. Murty, S. Murty,

and K. Dhavala, 2008: Counting the cost of vulture decline

– an appraisal of the human health and other benefits of

vultures in India. Ecol. Econ. 67, 194–204.

Table 1. Assessment of pairwise linkages of factors with respect to Nipah virus. For each pair, the factor listed in the left column is the primary

driver of the linkage, and the factor in the top row is the factor that may be affected (adapted from Gale and Jansen, 2010)

Affected factor

Migration Level of containment Temperature Conflict Host behaviour Poverty Deforestation

Primary Driver

Migration + NL +/? + ++ +

Level of containment NL NL NL + ? NL

Temperature ++ + + + + ++

Conflict + + NL + + +

Host behaviour NL + NL NL/+ + NL

Poverty + ++ NL + + +

Deforestation ++ + + + + +

NL, No apparent linkage; +, rational linkage.

P. Gale and A. C. Breed Horizon Scanning for Viruses

ª 2012 Crown copyright • Transboundary and Emerging Diseases. 3