The spread process of epidemic influenza in the continental United States, 1968–2008

https://doi.org/10.1016/j.sste.2014.01.001Get rights and content

Highlights

  • This paper modeled the spread process of epidemic influenza in the United States.

  • I examined spatial and population-based clustering at different times in a season.

  • Spatial clustering was only important at the peak of annual influenza waves.

  • Certain sections of the US had significant interstate spread throughout a wave.

Abstract

Understanding the quantitative disease dynamics of influenza is important in developing strategies to control its spread. This research analyzed the dominant spread process of epidemic influenza in the continental United States over a 41-year period. Spatial autocorrelation and simple correlation were applied to pneumonia and influenza mortality to observe the effect of distance and population on the between-state transmission of seasonal influenza. Annual influenza epidemics exhibited distance-based spatial spread at the peak of activity, but did not undergo significant population-based spread at any point. Geographically-close states (<500 miles) showed higher correlations in the start, peak and end of annual epidemics compared with geographically-distant states. Additionally, significant local clustering was found in the Midwest, Ohio River Valley and Northeastern regions as well as Nevada and Utah throughout an influenza season. This research may be combined with others in order to determine the main epidemic pathways of seasonal influenza in the US.

Introduction

Recurrent epidemics of influenza occur annually during the winter season in temperate areas of the world, such as the United States. They are the result of genetic drift in which, in order to escape host immunity, the surface antigens of influenza viruses undergo small changes (Smith et al., 2004). These annual influenza epidemics cause considerable morbidity, mortality and economic burden (Simonsen et al., 1999). In the US alone, approximately 24,000 deaths a year can be attributed to influenza (CDC, 2010).

It is crucial to understand quantitatively how a disease spreads in modern society. The sudden appearance of the 2009 H1N1 pandemic increased interest in the design of efficient containment policies and demands for an accurate characterization of spatial and temporal epidemic influenza patterns (Mills, 2006, Colizza et al., 2007, Ferguson et al., 2005, Ferguson et al., 2006, Longini et al., 2005). One of the most important control strategies that arose out of the 2009 pandemic was the need to identify the main channels of transmission or “epidemic pathways” of seasonal influenza in the US. In fact, identification of these pathways is the first clue on how to control influenza’s spread (Colizza et al., 2006).

While much is understood about the make-up and impact of seasonal influenza in the US, the spatial pattern of epidemic influenza has been less well characterized. Prior studies have analyzed the spread of influenza, developed unique models to describe and understand this disease, and explained the spatial distribution of influenza spread and of annual waves of infection in the US (Anderson and May, 1991, Baroyan et al., 1969, Bonabeau et al., 1998, Rvachev and Longini, 1985, Viboud et al., 2004). However, these studies have failed to detect the preferred channels of transmission for epidemic influenza in the US.

This study describes a method used to identify the dominant spreading process in the US. Spatial autocorrelation was combined with simple correlation to illustrate the disease dynamics of epidemic influenza. The objective of this study was to find robust transmission channels for epidemic spread by determining the dominant spread process of epidemic influenza in the continental United States over a 41-year period.

Section snippets

Methods

To examine the spatial structure of influenza epidemics in the continental United States between 1968 and 2008, monthly counts of pneumonia and influenza (P&I) mortality were obtained from the National Centers for Health Statistics (NCHS) and Centers for Disease Control and Prevention (CDC) Wonder Online Data System for Multiple Cause of Death (National Center for Health Statistics, 2012a, CDC, 2010). Population estimates used in the calculation of mortality rates were obtained from the Census

Results

Fig. 1 plots the monthly values of the spatial autocorrelation coefficient for the contagious (1A), hierarchical (1B) and mixed contagious-hierarchical (1C) diffusion graphs. These graphs also contain a plot of the monthly series of P&I mortality rate for the continental United States from January 1968 to December 2008. A vertical dashed line is set at z = 1.96 to indicate statistically significant I coefficients at the p = 0.05 level in a two-tailed test for positive spatial autocorrelation. A

Discussion

The spread of influenza in the continental US was characterized by reliance on distance and close proximity. A clearly defined process of spatial contagion drove the spread of annual influenza waves. Hierarchical diffusion played a much smaller role in influenza spread in the US and did not appear to influence spread during any part of an annual epidemic wave. Although the strength and timing of contagious transmission did vary between the waves, the broad findings implicate that spatial

Conclusion

Results of this research indicate that spatial proximity plays an important part in the spread process of seasonal influenza in the United States, particularly at the height of a wave. Distance-dependent spread may not be present in the US as a whole over an epidemic cycle, however, many parts of the country rely on interstate transmission between neighboring states to not only kickstart their annual influenza waves, but to sustain them through to termination. Spatial proximity may be a new and

Acknowledgments

The author would like to thank Dr. Stephen S. Morse for providing assistance with editing and review of the manuscript. The author would also like to note that this work was undertaken through the support of an Agency for Healthcare Research and Quality Dissertation Award (Grant #1R36HS021085-01).

References (25)

  • L. Simonsen et al.

    The global impact of influenza on morbidity and mortality

    Vaccine

    (1999)
  • R.M. Anderson et al.

    Infectious diseases of humans: dynamics and control

    (1991)
  • L. Anselin

    Local indicators of spatial association – LISA

    Geog Anal

    (1995)
  • O.V. Baroyan et al.

    An attempt at large-scale influenza epidemic modeling by means of a computer

    Bull Int Epidemiol Assoc

    (1969)
  • E. Bonabeau et al.

    The geographical spread of influenza

    Proc Biol Sci

    (1998)
  • J.S. Brownstein et al.

    Emperical evidence for the effect of airline travel on inter-regional influenza spread in the United States

    PLoS Med

    (2006)
  • Estimates of deaths associated with seasonal influenza – United States, 1976–2007

    Morb Motal Wkly Rep

    (2010)
  • V. Colizza et al.

    The role of airline transportation network in the prediction and predictability of global epidemics

    Proc Natl Acad Sci USA

    (2006)
  • V. Colizza et al.

    Modeling the worldwide spread of pandemic influenza: baseline case and containment interventions

    PLoS Med

    (2007)
  • P. Crépey et al.

    Detecting robust patterns in the spread of epidemics: a case study of influenza in the United States and France

    Am J Epidemol

    (2007)
  • N.M. Ferguson et al.

    Strategies for containing an emerging influenza pandemic in Southeast Asia

    Nature

    (2005)
  • N.M. Ferguson et al.

    Strategies for mitigating an influenza pandemic

    Nature

    (2006)
  • 1

    Present address: 3311 Toledo Road, Rm. 7222, Hyattsville, MD 20782, USA. Tel.: +1 301 458 4467.

    View full text