A Blog by Jonathan Low


Jan 9, 2017

Why Gun Shot Wounds Are Contagious: Data Reveals Spread Is Viral

Using data to expose the virality of network effects rather than relying on antiquated traditional notions. JL

Beth Mole reports in ars technica:

With disease modeling, scientists zero in on how gun violence spreads in Chicago. Disease models were the best at predicting gunshot data. That “contagion” model explained 63 percent of shootings. Gun violence transmission has an “incubation” period. The more time beyond that, the more likely that an “exposed” person wouldn’t catch gun violence illness. Modeling gun violence like disease could point to specific networks and individuals in need of intervention.
With our news reports speaking of gunfire epidemics, outbreaks, and plagues, firearm violence often sounds like a disease. But according to a new study, it often acts like one, too. In fact, catching a bullet may be a little like catching a cold—albeit a really bad one.
Gun violence can ripple through social networks and communities just like an infectious germ, Harvard and Yale researchers reported in JAMA Internal Medicine. This may not seem surprising, because earlier work has found that gun violence often clusters in certain areas and groups, particularly those steeped in gangs and drugs. But this study is the first to show that gun violence spreads directly from person to person after shootings—it’s not just about growing up in the same rough neighborhood or having the same risk factors.
The finding is good news, because, after decades of research, scientists are pretty good at predicting how infections cascade through populations. Applying disease-based theories and simulations to gun violence could help health workers get ahead of bullets and intervene before violence spreads. A more informed strategy could also cut down on intervention tactics that “rest largely on geographic or group-based policing efforts that tend to disproportionately affect disadvantaged minority communities,” the authors argue.
For the study, the researchers, led by sociologist Andrew Papachristos at Yale, dug into gunshot and arrest data from Chicago, covering the years between 2006 and 2014. The Windy City, which made headlines over the recent holidays for having more than 100 gunshot victims, is much like other big US cities in that its gun violence is intensely concentrated in specific neighborhoods.
Peering into arrest records, the researchers focused on 138,163 people that they considered part of a large social network, which the researchers based on people arrested together (the network represented 29.9 percent of those arrested in Chicago during the study period). Of those, 9,773 people were involved in 11,123 gun violence incidents, both fatal and nonfatal.
Next the researchers ran different models of how gun violence spread through the network over time, taking into account demographic data such as sex, age, and neighborhood. They also included the data on who was arrested together—partners in crime are often known to have close social ties.
Comparing a model based just on demographics with one that mimicked infectious disease simulations, the disease models were the best at predicting the gunshot data. That “contagion” model explained 63 percent of the shootings.
The analysis also suggested that gun violence transmission has an “incubation” period—which, in terms of disease, is the gap between being sneezed on and when you actually get sick. The researchers found an average of 125 days between when an individual experienced gun violence and when their “infector” experienced gun violence.
The more time beyond that 125-day period, the more likely it was that an “exposed” person wouldn’t actually catch the gun violence illness. Researchers didn’t have enough data to figure out why some people didn’t catch the disease and others did—it’s a question for the next study, they suggested. They also don’t know how well the infectious disease modeling will work in other cities, which have different gun laws, gang activity, public housing policies, segregation, and more.
Nevertheless, the researchers think that modeling gun violence like disease outbreaks could point heath workers to specific networks and individuals in need of intervention.
In a separate commentary piece, epidemiologists Charles Branas, Sara Jacoby, and Elena Andreyeva of the University of Pennsylvania praised the study. “This important finding helps put to rest the mistaken idea that epidemiology, medicine, and public health somehow have no place in the prevention of firearm violence, a disease process that affects roughly 100,000 people in the United States each year.” However, they noted that we still shouldn’t abandon efforts and interventions to improve neighborhood-wide conditions, which are also known to help curb gun violence.


Post a Comment