Science Life

Cell Phone Data Creates a Tool for Stopping HIV

A bus stop in India, one of the “hotspots” used by John Schneider and his team to recruit interview subjects for the study

For much of the past decade, epidemiologist John Schneider has spent time in India, analyzing social networks of populations at risk for contracting HIV, to better understand interpersonal relationships and identify key members of the community. His goal is find people with the connections and credibility to educate others about safe sex practices.

In a recent paper published in the journal Social Science & Medicine, he and colleagues show how they were able to use a ubiquitous feature of modern life – the cell phone – to create a more accurate picture of the social networks that play a central role in the spread of HIV in India.

His team has been able to persuade many high-risk community members to let them extract contact list data from the SIM cards of their cell phones, and then used it to build objective representations of their social networks. With those models, they could then identify the people who might be the most effective at educating their peers and potentially limiting the spread of HIV.

In 2011, Schneider was conducting fieldwork in Hyderabad, a large city in southern India, interviewing truck drivers about their networks and sexual behaviors. Truck drivers are one of the populations at highest risk for contracting HIV in India. Many drivers solicit sex from each other or with paid sex workers at truck stops and social venues along their routes. Because these drivers travel far and wide across the country, they are central to the spread of the disease and thus a crucial link to stopping it.

John Schneider (right) with A. Ning Zhou, Pritzker School of Medicine MD candidate and co-author of the study

During these interviews, Schneider said, he and his research assistant would ask how many friends each driver had. Drivers would give one or two names and claim they had few friends, but at the same time the ringing of their cell phones constantly interrupted the interviews. The tone of these phone conversations made it obvious that these weren’t just calls from random people, so when pressed about who was calling, the drivers allowed that they were other friends they’d forgotten to mention.

Schneider said at this point he realized that data from these cell phones could be a much better source of information about social networks than relying on the drivers, who may forget people or omit key relationships for various reasons.

“Analyzing the cell phones provides a lot of objective information,” he said. “If I asked you now to tell me all your friends, you’d be able to remember a fair amount of them. But if I went into your cell phone, then I’d have information about your family, different acquaintances, business contacts. So getting in the cell phones really provides full, rich information.”

They analyzed the networks built from this data to find the most highly-connected and influential members. They then conducted follow up interviews about these individuals’ sexual behaviors and social characteristics, such as how open they are to new ideas, that might give clues to their potential to educate their peers.

This method of intervention is called the peer-change-agent model. A peer-change agent is someone in the community trained to educate others. They’re often chosen by subjective measures like popularity or charisma, or whether they’re considered an opinion leader in the network.

But Schneider said these characteristics don’t always make a person the best change agent. A popular person might be too busy, or may not want to risk his popularity by broaching sensitive subjects like circumcision or condom use. Sex workers are often recruited to spread prevention education, but the dynamics of their client relationships may limit their influence. What’s more important, Schneider said, is where they’re located in the network.

“We can train anyone to provide messages, recruit people and interact with their social network, but we think that it’s really where they are in the network that’s also important,” he said. “That’s one facet of this that’s really been left out in previous studies of this model.”

Research assistant Sabitha conducts an interview with a driver

In this study, they found that people who serve as “bridges” between different parts of the network—people who travel regularly between geographic areas, or who connect different social groups—are just as likely to encourage condom use as trained peer educators, and less likely to engage in risky behavior such as accepting money for sex. They were also more likely to be innovators than those more centrally located in the group, i.e. willing to try out new ideas or encourage others to change their behavior.

Schneider said it’s to be expected that people at the edges who straddle different geographic areas or social groups are more open to new ideas than those highly connected in the center. But it’s the first time it’s been shown in a health context that being in a certain position in the network makes them more innovative.

The next step, he said, is to see if training these bridges in prevention can be an effective means of peer education. And Schneider said it’s not about just education either; it’s about stopping the spread of HIV itself.

“They are at key places to stop the transmission of HIV within networks,” he said. “We also want them to incorporate some of that knowledge and training so they’re protecting themselves, and the transmission pathways aren’t going through them.”

We want to hear from you! Take a brief survey to help us improve Science Life

=====
Schneider J.A., Zhou A.N. & Laumann E.O. (2014). A new HIV prevention network approach: Sociometric peer change agent selection, Social Science & Medicine, DOI: