Long-term dynamics of measles in London: Titrating the impact of wars, the 1918 pandemic, and vaccination

• Published in: PLoS computational biology 2019-09, Vol.15 (9), p.e1007305-e1007305
• Main Authors: Becker, Alexander D, Wesolowski, Amy, Bjørnstad, Ottar N, Grenfell, Bryan T
• Format: Article
• Language: English
• Subjects: Biology and Life Sciences; Computational Biology; Computer and Information Sciences; Demographic aspects; Demographics; Demography; Dynamical systems; Ecological monitoring; Ecology; Ecology and Environmental Sciences; Epidemics; Epidemiology; Evolutionary biology; Forecasts and trends; History, 20th Century; Humans; Immunization; Incidence; Infectious diseases; Influenza; Influenza vaccines; Influenza, Human - epidemiology; Influenza, Human - history; Internet; Life cycles; London - epidemiology; Markov analysis; Mathematical models; Measles; Measles - epidemiology; Measles - history; Measles - prevention & control; Measles - transmission; Medicine and Health Sciences; Mortality; Nonlinear dynamics; Nonlinear systems; Pandemics; Pandemics - history; Pandemics - statistics & numerical data; Phase transitions; Physical Sciences; Public health; Stochasticity; Time series; Vaccination; Vaccination - history; Vaccination - statistics & numerical data; World War I; World War II
• ISSN: 1553-7358; 1553-734X; 1553-7358
• DOI: 10.1371/journal.pcbi.1007305

A key question in ecology is the relative impact of internal nonlinear dynamics and external perturbations on the long-term trajectories of natural systems. Measles has been analyzed extensively as a paradigm for consumer-resource dynamics due to the oscillatory nature of the host-pathogen life cycle, the abundance of rich data to test theory, and public health relevance. The dynamics of measles in London, in particular, has acted as a prototypical test bed for such analysis using incidence data from the pre-vaccination era (1944-1967). However, during this timeframe there were few external large-scale perturbations, limiting an assessment of the relative impact of internal and extra demographic perturbations to the host population. Here, we extended the previous London analyses to include nearly a century of data that also contains four major demographic changes: the First and Second World Wars, the 1918 influenza pandemic, and the start of a measles mass vaccination program. By combining mortality and incidence data using particle filtering methods, we show that a simple stochastic epidemic model, with minimal historical specifications, can capture the nearly 100 years of dynamics including changes caused by each of the major perturbations. We show that the majority of dynamic changes are explainable by the internal nonlinear dynamics of the system, tuned by demographic changes. In addition, the 1918 influenza pandemic and World War II acted as extra perturbations to this basic epidemic oscillator. Our analysis underlines that long-term ecological and epidemiological dynamics can follow very simple rules, even in a non-stationary population subject to significant perturbations and major secular changes.