system dynamics  
At Any Rate

by Chris Soderquist and Bill Harris

Model 3
, February 2002

When Smallpox Becomes a Threat Again

One of the legacies of September 11 is a heightened awareness of the fragility of human life and our susceptibility to bioterrorism. Smallpox, declared eradicated over 20 years ago, is in the news again, forcing policymakers to make many difficult decisions. Should they restart the vaccinations that were part of everyone's life decades ago? Should they focus on stockpiling vaccines? Should they rely on quarantines? How should we as citizens evaluate what the government and health officials decide?

We believe systems thinking can offer insight into these questions to those people responsible for dealing with such horrific events as a smallpox outbreak. We've developed a model structure that generates the behavior experienced in an epidemic. This model can help us to determine policies that might impact outcomes as well as describe potential consequences, both intended and unintended, of these policies. It also allows us to explore the issue publicly, and to articulate, test, and debate different opinions and assumptions clearly. Such an approach won't give the answer, but it can provide structure to a forum that seeks to improve confidence in policy decisions.

Let's see how systems thinking might assist government and public health officials in dealing effectively with a smallpox epidemic and help citizens evaluate the work of their government. We'll start, as we often do, by describing the behavior we're trying to understand: What's the basic characteristic of an epidemic? Many people getting sick very suddenly. It starts out with one or two victims, then the number escalates rapidly. After a while, the epidemic runs its course, and the incidence of illness subsides almost as quickly as it rose. The pattern of people with the disease looks like the graph below.
.


Next, let's create a hypothesis about the nature of a structure that can generate the behavior shown in the graph. First, it has a stock of sick people. Because these people started out as healthy, we'll consider healthy people a stock, too. (Remember, stocks are accumulations-they're like bathtubs to which water is added or taken away. They are shown as rectangles in our model diagram). We'll consider the number of people becoming infected per day as a flow between the stocks of healthy people and sick people. (Flows are what cause the stock to change; they add and take away what's in the stock. They are shown as pipes with valves.)

We know, of course, that people don't stay sick forever. Something happens: either they eventually recover or, unfortunately for some, they die. If they recover, they are typically immune from further contracting of that disease. The four main stocks in our model-healthy people, sick people, recovered people, and deceased people-and the three flows connecting them describe how people "move" through an epidemic. By including these variables, we have made the model complete enough to recreate the evolution of an epidemic-in this case, the rapid increase in small pox occurrence, followed by a later, rapid decline in occurrence.




How might we use such a model to test strategies for dealing with an epidemic? Public health and government officials, who must decide how to prepare for the eventuality of bioterrorism attacks, can use such a model:

To compare the effectiveness of immunization and quarantine programs and determine how good such programs would have to be.
To evaluate if either program alone is sufficient to address the epidemic or if both are needed.
To explore possible adverse consequences of vaccinations.
To share their decisions with the public to improve people's understanding of their recommendations, thus creating more public support (assuming their policies appear good!).

Below we describe two possible public policies, one based on quarantines, the other immunizations, which the model can test for efficacy.

Policy 1: Quarantines
Perhaps the oldest approach to dealing with highly infectious diseases, a quarantine involves segregating those who are infectious from the rest of the population. In the case of a smallpox attack, public health systems would likely be called into action the moment a case was identified. As people were diagnosed with the disease, they and other people they may have infected would be isolated. Unfortunately, in most cases people would become infectious at the same time that they began to exhibit symptoms, and thus they'd have potentially contacted a few others before they were isolated. Quarantines, if conducted too aggressively and without sufficient reason, can generate legal challenges.

We've designed the downloadable model to initiate quarantines after an outbreak has been detected. The user can define how long it takes to initiate a quarantine and how effective it is. The model indicates that a quarantine must be highly effective to have a significant impact. At a 50 percent effectiveness rate, the primary effect of the quarantine is to slow the growth of the disease. The total number deceased doesn't change much, at least in the six-month time horizon of the model. As the quarantine improves to filter out more than 90 percent of infected people, the number who die in the first six months declines markedly.

Policy 2: Immunizations
Immunizations involve preventing the disease from occurring, rather than stopping its spread, but a vaccine must exist, and it must be available in sufficient quantity. The vaccine also has a risk of side effects, which can include death.

Similar to quarantines, the model simulates an immunization program by sensing an outbreak of smallpox, waiting for the delay in starting the vaccination program, and then reducing healthy people's susceptibility to the disease. The user can vary the time to initiate immunizations and vary the effectiveness of that immunization program. Since an immunization program prevents people from catching the disease, while the quarantine inhibits its further spread, a good immunization program has advantages over a good quarantine.

The individual results for quarantines and immunizations suggest that a combination strategy would be effective: quarantine people immediately to slow the epidemic, and vaccinate those likely to come in contact with known infected people as quickly as you can. While our model doesn't simulate this combined approach, if you have a copy of ithink® or STELLA® software, you can modify the model to combine quarantines and immunizations.

We invite you to download the model now to learn more about how epidemics might evolve and the possible results of quarantine and immunization policies. Think about some of the assumptions we didn't model. What do you like in the model? What do you think should be changed? Let us know your thoughts in the Pegasus forum. We realize that public health officials would need to develop and test the model more thoroughly before using it to set public policy. Our goal in designing it was to illuminate how systems thinking can improve the richness and rigor of discussions concerning possible approaches to a bioterrorism attack.

The smallpox model was created in collaboration with Turning Point's National Program Office, which is funded by the Robert Wood Johnson Foundation.

 

Using the Model
To use the model, you'll need to download two files—the "current model" and the "isee Player" (the ithink® Runtime for the At Any Rate model series) that runs the model. Both are located in the "Get" section toward the top of the right-hand column. You'll then need to install the isee Player on your computer. (Once you have installed the isee Player on your computer, you no longer have to go through this process unless the reader is updated.)

1) Download the "Current Model"
• Click "Current Model."
• Choose "Save this file to a disk" and click "okay."
• In "Save As," save the ITR file to your desktop (or to a folder of your choosing).

2) Download and install the "isee Player"
• Follow the instructions on the isee Systems site.
After you install the isee Player, to run the model, you can go to your desktop and double-click on "model1.itr" or start the ithink® program and use the "file open" command to locate and open the model1.itr file.

You are ready to begin. Feel free to play with the model. We've put more content in it than we've described in this column. Try different things. If you've got an interesting idea, a question, or a comment, go to our Pegasus Forum. We'd enjoy hearing from you.
 

This learning lab was developed using the ithink® software, a computer simulation modeling package developed and distributed by isee Systems.

 

About
At Any Rate 
Bill Harris
iThink software

 
Get

  This model
  isee Player
   

Dialogue

At Any Rate Pegasus forum
 

Related links
Smallpox Home Page, provided by the U.S. Centers for Disease Control 
Smallpox information, provided by the World Health Organization
John D. Sterman, "S-Shaped Growth: Epidemics, Innovation Diffusion, and the Growth Of New Products" in Business Dynamics: Systems Thinking and Modeling for a Complex World (McGraw-Hill, 2000), chapter 9. 
Turning Point


Resources
Systems Thinking
System Dynamics
Causal Loop Diagrams
Systems Archetypes



Archive

Feedback
Send comments to our editor



The Gateway
ConferencesNewslettersProduct GalleryLearn MoreAbout PegasusGuestbookHome
Audio & VideoBooksLearning PackagesSoftware & Games Visual Tools



Search for Products ConferencesThe Systems ThinkerLeverage PointsBulletin Boards



Shopping Cart How to OrderSearch & OrderHelpFAQSite Map


All materials © 2000-2006 Pegasus Communications, Inc.
Phone: 1-781-398-9700• 1-800-272-0945 Fax: 1-781-894-7175
Mail:1 Moody Street, Waltham, MA 02453-5339
Customer Service: customerservice@pegasuscom.com
Editoral/Business Offices: info@pegasuscom.com