Strategies for Controlling Non-Transmissible Infection Outbreaks Using a Large Human Movement Data Set

Penelope A. Hancock*, Yasmin Rehman, Ian Hall, Obaghe Edeghere, Leon Danon, Thomas A. House, Matthew J. Keeling

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

4 Citations (Scopus)


Prediction and control of the spread of infectious disease in human populations benefits greatly from our growing capacity to quantify human movement behavior. Here we develop a mathematical model for non-transmissible infections contracted from a localized environmental source, informed by a detailed description of movement patterns of the population of Great Britain. The model is applied to outbreaks of Legionnaires' disease, a potentially life-threatening form of pneumonia caused by the bacteria Legionella pneumophilia. We use case-report data from three recent outbreaks that have occurred in Great Britain where the source has already been identified by public health agencies. We first demonstrate that the amount of individual-level heterogeneity incorporated in the movement data greatly influences our ability to predict the source location. The most accurate predictions were obtained using reported travel histories to describe movements of infected individuals, but using detailed simulation models to estimate movement patterns offers an effective fast alternative. Secondly, once the source is identified, we show that our model can be used to accurately determine the population likely to have been exposed to the pathogen, and hence predict the residential locations of infected individuals. The results give rise to an effective control strategy that can be implemented rapidly in response to an outbreak.

Original languageEnglish
JournalPLoS Computational Biology
Issue number9
Publication statusPublished - 1 Sept 2014

Bibliographical note

Publisher Copyright:
© 2014 Hancock et al.


Dive into the research topics of 'Strategies for Controlling Non-Transmissible Infection Outbreaks Using a Large Human Movement Data Set'. Together they form a unique fingerprint.

Cite this