Simulation modeling for pandemic decision making: A case study with bi-criteria analysis on school closures

https://doi.org/10.1016/j.dss.2012.10.013Get rights and content

Abstract

Pandemic influenza continues to be a national and international public health concern, and has received significant attention worldwide with the A/H1N1 influenza outbreak in 2009. Many countries, including the United States, have developed preparedness plans for an influenza pandemic. Preparedness plans are falling under renewed scrutiny as decision-makers apply new findings and seek key leverage points for more effective preparedness and response. School closure has been recommended by the World Health Organization as one of the best ways to protect children and other susceptible individuals at the early stages of the pandemic. However, school closure is a difficult mitigation policy to implement from both strategic and operational points of view. Challenges include impacts on alternative education delivery services, such as student meals and after-school oversight, as well as direct and indirect economic outfalls. To help public health decision makers address these issues, we developed an epidemiological simulation tool for pandemic influenza which enables users to make decisions during a simulated pandemic. We then designed a school closure tabletop exercise using our simulation model as a decision-support tool for evaluating the effectiveness of school closure as a community mitigation strategy for pandemic influenza. We conducted two exercises in February 2009 for the Arizona Department of Health and Human Services including high-ranking health and education administrators from across the state. The purpose of these exercises was to test the state's pandemic preparedness plans with respect to school closure timing and impact. The exercises required participants to make (hypothetical) strategic and operational decisions to mitigate the impacts of pandemic influenza at the state and local levels. Our simulation and decision analysis tool was used to assess the impact of key decisions in the exercises. This paper presents the technical details involved in the design and evaluation of this pandemic decision-support tool. Based on the decisions made in the exercises, we present a bi-criteria decision analysis framework to evaluate analytical results obtained from the simulation model. Our analyses show that sequential school closure and re-opening strategy with a specific decision rule gives the best compromised solution in terms of minimizing the total number of infections and providing minimal educational discontinuity.

Introduction

In 2009, the influenza A/H1N1 virus quickly escalated to a global pandemic with sustained human-to-human transmission worldwide [13], [41]. This recent influenza outbreak highlighted the importance of preparedness activities for all countries in today's global world. Current preparedness efforts for pandemic influenza focus on establishing efficient mitigation strategies for communities and planning for delivery of necessary medical services during an outbreak. School closure is a non-pharmaceutical influenza mitigation strategy and has been included in the federal guidelines for the Community Strategy for Pandemic Influenza Mitigation [8]. During the 2009 A/H1N1 influenza outbreak, local authorities across the United States closed schools in an effort to decrease influenza transmission rates. However, there is limited quantitative and scientific evidence available to support the efficacy of this influenza mitigation measure, which has profound legal, economic, and social implications.

Tabletop exercises are used by many agencies and organizations in preparation for various disasters. A tabletop exercise is an informed discussion by responsible authorities for testing the theoretical and hypothetical capabilities in response to a disaster, such as pandemic influenza [3]. These exercises are bounded by a scenario or set of scenarios and include public health professionals and policy-makers responding to prompted discussion questions. Although tabletop exercises for pandemic influenza preparedness have not typically included computer-based simulation models, decision support systems or multi-media enabled narrative and real-time feedback systems [7], [36], these tools are becoming increasingly important for public health managers for surveillance, prediction and decision analysis [44]. Previous work by Yoon et al. has shown that computer based virtual systems can significantly improve the ability of trainees to respond to dynamic and chaotic environments interactively and collaboratively [43].

In an attempt to test existing influenza preparedness plans in Arizona, the Arizona Department of Health and Human Services (ADHS) hosted two tabletop exercises at the Arizona State University Decision Theater on February 12 and February 19, 2009. The Decision Theater is a visualization center where a collaborative decision-making environment is supported with various technological tools such as stereo sources, geospatial visualizations and dynamic simulation models. The participants of these exercises included key individuals who are involved in school closure decisions in Arizona (the director of the largest county health department in Arizona, several school superintendents, and Arizona Department of Education representatives). Participants also included school nurses, first responders and state, county, and local health officials as well as members of Parent Teacher Associations, large community businesses, and policy makers and planners at the K–12 and university levels.

The 2009 influenza tabletop exercises at the Decision Theater focused primarily on school closure decisions for pandemic mitigation. These exercises integrated traditional elements of a tabletop design with a computer-based epidemiological simulation model, science-based narratives using mock news videos, a web-based participant collaboration and discussion-feedback tool, and facilitated group discussion. Participants of the exercises were prompted to make critical decisions (i.e. when to close schools, how long to close and when to re-open them), which impacted the results of the model.

State health officials benefited from the simulation model by defining and testing key epidemiological metrics for school closure identified in preparedness plans. The simulation model allowed decision makers to evaluate alternative strategies for making school closure decisions (including duration of closure, timing of closure and number of schools closed) which helped prepare them for making critical school closure decisions during the A/H1N1 outbreak just a few months later. In this paper, we outline the organizational design of the school closure exercises, and explain the technical specifications of the supporting simulation–visualization tool. We then present the results generated by the simulation model and analyze them in a bi-criteria evaluation framework. In addition, this paper outlines a case in which a simulation–visualization based decision support tool can be effectively used for practicing policy decision making processes.

Section snippets

Literature review

Previous research in the field of pandemic influenza modeling is quite extensive and can be broadly categorized as: 1) modeling the spread of pandemic in a community [4], [6], and building geospatial and temporal simulation models to estimate the global path of the pandemic spread [11], [17], [26], [30], 2) evaluating possible non-pharmaceutical and pharmaceutical interventions at the global and local levels [10], [14], [42] and recently, 3) logistical issues related to the community response

System design and integration for dynamic decision making

For pandemic preparedness exercises, we have taken a systems approach and first developed a geospatial and age structured simulation model of influenza spread in Powersim, a system dynamics modeling and simulation software [29]. This model is fed with several parameters which are statically stored in an excel spreadsheet. The simulation model is also connected to a central database where several inputs and resulting outputs from the model are kept in different forms (spreadsheets, raster etc.).

Case study

This section explains the details of the exercises that took place in February 2009 at the Decision Theater of Arizona State University. The developed simulation–visualization tool was used to facilitate and coordinate these exercises for the Arizona Department of Health Services (ADHS). The exercise participants tested their pandemic preparedness plans with respect to school closure timing and impacts.

Conclusions

As we have seen with the recent H1N1 outbreak, the education system could be severely strained during a pandemic. Planning at all levels of response from the federal level down to individual schools is required to meet public health challenges faced by a pandemic. Additionally, roles and responsibilities for all responders and stakeholders must be clearly delineated in response plans, and coordinated exercises must be conducted to prepare our communities, schools, and children for a prolonged

Acknowledgments

We are thankful to the Arizona Department of Health and Human Services and the Public Health Department of Maricopa County for their support and collaboration on the design of the presented exercises and sharing their pandemic preparedness plans with us. We also acknowledge the valuable comments from Dr. Bob England, Director of Public Health Department in Maricopa County during and after the exercises. We acknowledge the help from Dr. Robert Pahle of Decision Theater for his help to generate

Ozgur Araz is an assistant professor in the College of Public Health at the University of Nebraska Medical Center. His primary research interests are complex systems thinking, modeling and simulation, modeling chronic and infectious diseases, stochastic optimization, healthcare operations management and sustainable development. He has his Ph.D. in Industrial Engineering from the Arizona State University. His email address is [email protected].

References (44)

  • C. Castillo-Chavez et al.

    Epidemiological models with age structure, proportionate mixing and cross immunity

    Journal of Mathematical Biology

    (1989)
  • D.J. Dausey et al.

    Designing and conducting tabletop exercises to assess public health preparedness for manmade and naturally occurring biological threats

    BMC Public Health

    (2007)
  • Department of Health and Human Services (US)

    Community strategy for pandemic influenza mitigation. February 2007. [Cited 2011 May 11]

  • A. Ekici et al.

    Pandemic influenza response

  • 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)
  • A. Flahault et al.

    A mathematical model for the European spread of influenza

    European Journal of Epidemiology

    (1994)
  • C. Fraser et al.

    Pandemic potential of a strain of influenza A (H1N1): early findings

    Science

    (2009)
  • T.C. Germann et al.

    Mitigation strategies for pandemic influenza in the United States

    Proceedings of the National Academy of Sciences

    (2006)
  • M.Z. Gojovic et al.

    Modeling mitigation strategies for pandemic (H1N1) 2009

    CMAJ

    (2009)
  • R.F. Grais et al.

    Assessing the impact of airline travel on the geographic spread of pandemic influenza

    European Journal of Epidemiology

    (2003)
  • R.F. Grais et al.

    Modeling the spread of annual influenza epidemics in the US: the potential role of the air travel

    Healthcare Management Science

    (2004)
  • Cited by (0)

    Ozgur Araz is an assistant professor in the College of Public Health at the University of Nebraska Medical Center. His primary research interests are complex systems thinking, modeling and simulation, modeling chronic and infectious diseases, stochastic optimization, healthcare operations management and sustainable development. He has his Ph.D. in Industrial Engineering from the Arizona State University. His email address is [email protected].

    Tim Lant is an assistant research professor at the ASU Decision Theater. His research interests are on public health emergency preparedness for emerging pandemics and bioterrorism, public health policy-making, decision-support, analytics, visualization, and simulated emergency preparedness exercises. He earned his Ph.D. in applied mathematics with specialization in mathematical biology at Arizona State University. His email address is [email protected].

    John Fowler is the Motorola Professor of International Business and Chair of the Supply Chain Management Department in the WP Carey School of Business at Arizona State University (ASU) and a professor of industrial engineering in the School of Computing, Informatics, and Decision Systems Engineering at ASU. His research interests include modeling, analysis, and control of manufacturing and service systems. Dr. Fowler is the founding editor of the IIE Transactions on Healthcare Systems Engineering and is a joint editor of the Journal of Simulation. His email address is [email protected].

    Megan Jehn is an associate professor in the School of Health Management and Policy at Arizona State University. Her research interests are public health emergency preparedness and health policy. She has her Ph.D. in epidemiology from the Johns Hopkins Bloomberg School of Public Health. Her email address is [email protected].

    View full text