Epidemics and population structure: One step forward, and two steps back

Duration: 30 mins 52 secs

Share this media item:

Embed this media item:

<iframe width="_width_" height="_height_" src="https://sms.cam.ac.uk/media/1542722/embed" frameborder="0" scrolling="no" allowfullscreen></iframe>

Choose size:

About this item

Available Formats

About this item

Description:	House, T (University of Warwick) Tuesday 20 August 2013, 09:00-09:30


Created:	2013-08-22 14:54
Collection:	Infectious Disease Dynamics
Publisher:	Isaac Newton Institute
Copyright:	House, T
Language:	eng (English)


Abstract:	In general, the incorporation of population structure into epidemic models creates problems of dimensionality for prediction (the `forward problem'). Even for the `simple epidemic' / SI model, complete individual heterogeneity of n individuals leads to a dynamical system whose size grows like 2^n. There are, however, two `inverse problems' where this curse becomes a blessing: for statistical inference, flat directions in parameter space can become identifiable once more stratification of data is available; and the presence of population structure allows a far wider range of control and mitigation strategies to be compared than are possible in a homogeneous system. This talk will consider: (i) the generation of predictions from heterogeneous epidemic models without excessive dimensionality; (ii) the use of multiple stratified data sources to resolve statistical questions about the otherwise unidentifiable but epidemiologically important quantities; (iii) informing public health policy on the basis of these considerations. Real-world examples will come from the 2009 H1N1 influenza pandemic.

Abstract:

In general, the incorporation of population structure into epidemic models creates problems of dimensionality for prediction (the `forward problem'). Even for the `simple epidemic' / SI model, complete individual heterogeneity of n individuals leads to a dynamical system whose size grows like 2^n.

There are, however, two `inverse problems' where this curse becomes a blessing: for statistical inference, flat directions in parameter space can become identifiable once more stratification of data is available; and the presence of population structure allows a far wider range of control and mitigation strategies to be compared than are possible in a homogeneous system.

This talk will consider: (i) the generation of predictions from heterogeneous epidemic models without excessive dimensionality; (ii) the use of multiple stratified data sources to resolve statistical questions about the otherwise unidentifiable but epidemiologically important quantities; (iii) informing public health policy on the basis of these considerations. Real-world examples will come from the 2009 H1N1 influenza pandemic.

Available Formats

Format	Quality	Bitrate	Size
MPEG-4 Video	640x360	1.94 Mbits/sec	449.40 MB	View	Download
WebM	640x360	1.23 Mbits/sec	286.74 MB	View	Download
iPod Video	480x270	522.13 kbits/sec	118.04 MB	View	Download
MP3	44100 Hz	249.78 kbits/sec	56.53 MB	Listen	Download
Auto *	(Allows browser to choose a format it supports)