Infographic: Predicting The Spread Of HIV, With A Virtual Community Of 150,000

Real data about sex and drugs fuses with supercomputing to simulate the spread of HIV between real (virtual) people.

Infographic: Predicting The Spread Of HIV, With A Virtual Community Of 150,000

Without a cure for HIV, its spread is inevitable. We have a finite number of dollars to fight it with any number of potential policies. And if we spend those dollars on poorly designed prevention plans, there’s no do-over. More people will get HIV, and that’s that.


Brandon Marshall is an assistant professor of epidemiology at Brown University. He’s created a computer simulation of NYC that can measure the spread of HIV in the face of various health care interventions–with what looks to be incredible accuracy. Why trust his model? It’s already replicated New York’s HIV infection rates between 1992 and 2002.

“In the absence of a cure or vaccine for HIV, one of the most promising strategies is to use a combination of evidence-based prevention programs to stop its spread. However, it is very challenging and expensive to design a clinical trial to identify which combinations of programs would be most effective,” Marshall tells Co.Design. “With a mathematical simulation, we can predict the success of a whole variety of combination prevention scenarios much faster and cheaper.”

Marshall starts with data straight from New York City–demographic information like drug use, sexual preference, needle sharing, viral transmission, and access to drug treatment. Then he maps these scenarios to 150,000 “agents,” or highly detailed individuals who make no promise to live saintly lives in the virtual world. They may use drugs and have unprotected sex with one another, or they may live life without a single close call of HIV. It’s all decided by the supercomputer that crunches the numbers as stories play out.

“With this level of detail, we can identify specific agents who are responsible for a larger number of new infections,” Marshall explains. “We can also examine how the sexual and injecting networks of high-risk groups affects the overall HIV epidemic in the general population.”

It really can come down to the role of just one person–a single person public policies need to reach–and we can see just what that means in an example from Brown’s official press release:

In one run, agent 89,425, who is male and has sex with men, could end up injecting drugs. He participates in needle exchanges, but according to the built-in probabilities, in year three he shares needles multiple times with another injection drug user with whom he is also having unprotected sex. In the last of those encounters, agent 89,425 becomes infected with HIV. In year four he starts participating in drug treatment and in year five he gets tested for HIV, starts antiretroviral treatment, and reduces the frequency with which he has unprotected sex. Because he always takes his HIV medications, he never transmits the virus further.

The result is probabilities played out in real scenarios, what are essentially micro lives measured on a macro scale. Then, to ensure accuracy of results, Marshall reruns the simulation thousands of times, rounding outliers as you would in any longitudinal study. But the real promise of such simulations isn’t to look at just one set of policies. It’s to look at many, in the most practical, life-saving terms possible.


“In a virtual system, we can tinker with different behaviors and other parameters, and predict the spread of HIV under a number of different scenarios,” Marshall explains. “In the near future, we are hoping to incorporate costing data to examine the cost-effectiveness of various program and policy scenarios. This is an important next step, since policymakers have to weigh the costs of each option with their expected impact.”

On one hand, it’s a cold, metricized way to weigh public policy–consider that we may not choose the most effectively simulated program because of costs, much like, in war, some soldiers don’t receive reinforcements in interest of the greater whole–on the other hand, we’ve been granted a virtual glimpse into the future, a means to make several mistakes before funding the most efficient plan that we can dream up. It’s approaching public policy with the scientific method, but without treating our own population as the lab rats.

In the future, Marshall is tweaking his HIV modeling program to model more specific environments. He’ll be analyzing how drug-using agents may spread the virus through our jail system. And while stopping the spread of HIV across all of NYC seems nearly impossible, setting up small, designed interventions within the confines of our prison system seems wholly possible.

[Hat tip: Medgadget]

About the author

Mark Wilson is a senior writer at Fast Company. He started, a simple way to give back every day.