A few years ago, a pair of sociologists named Andrew Papachristos and Christopher Wildeman decided to study gun violence in Chicago. They focused on a specific community on the west side: overwhelmingly black and disproportionately poor, with a murder rate that was five times higher than the rest of the city.

Their approach was to look at gun violence the way epidemiologists study disease — examining the way it spread by social connections. And like a virus, they found that there were certain people who were especially at risk of being touched by it.

"The one thing I found that was so surprising was the severe concentration of violence," Papachristos told me in 2012. The researchers isolated one particularly tight social cluster that was in danger of deadly violence — 4 percent of the people in that network accounted for almost 41 percent of the homicides in the neighborhood. They tended to be men. Almost all had a previous arrest. They tended to live about 350 feet from another homicide victim. And they tended to be only a couple of degrees removed socially from one another.

"They're a couple handshakes away from each other," Papachristos said. The further away someone in those communities was from that network, though, the safer he or she was, and the researchers argued that making those folks in the cluster more safe required a new approach. "Changing networks means changing communities," Papachristos later explained to the Science of Us. "You can't arrest your way out."

I remembered that conversation last week after the FBI released the national crime statistics for 2015. Although last year remained one of the safest years on record, the number of homicides jumped, and in a big way: killings were up nearly 11 percent over the previous year.

There are a bunch of ways to read that data. Violent crime had been historically low over the last half-decade and any uptick might look relatively dramatic. And 2015 might also turn out to be an aberration — criminologists say crime tends to fluctuate from year to year, and the next year's stats might again show a continuation of that general downward trend. Liberals and conservatives also tend to have very different interpretations of those numbers. But what's undeniable is that a major reason for the homicide surge is that at least 900 more black men and boys were killed in 2015. More than half of the jump in homicides could be attributed to spikes in in Chicago, Baltimore and Washington, D.C. It's not an overstatement to say that the phenomenon of surging American homicide is largely the problem of black men and boys.

Yet we treat them more as the potential perpetrators of violence and seldom as its most likely victims. Our approach hasn't been to identify and protect the people who might be most at-risk. Our approach has effectively been to treat everyone in those communities like a criminal. We've tried to arrest our way out of it.

In their efforts to reduce violence in places where it has remained stubbornly elevated, police departments in big cities have cycled through a host of strategies that put their officers in regular contact with black men —"hot-spot" policing, stop-and-frisk, biking while black, etc. That repeated exposure to the police results in routine, random searches of law-abiding people, yielding little more than thousands of summonses and arrests for petty infractions. Papachristos offered me a different idea: Rather than casting a wide net and indiscriminately ensnaring black men, we should focus much more on looking out for those most in danger — that is to say, a "hot-people" strategy instead of a "hot-spot" strategy.

But when the demographic skew in homicides is acknowledged, it tends to be unhelpfully labeled "black-on-black crime" — that is, murder as the inevitable result of a tangle of dysfunctional pathologies exhibited by black people. The proponents of this view say that any honest discussion about race and crime — and thus, race and policing — has to grapple with what they say is a deeply rooted flaw in the culture of black communities. And so we've dealt with this problem using the blunt instrument of policing, not imagination or even empathy.

Chicago, again, offers a vivid example of how that plays out. The city's deeply segregated black communities have long been subjected to ineffective and sometimes disastrously bad policing. Here are a few police scandals that became national news in just the last two years: there was the district where for decades, residents were routinelytortured and coerced by the police into giving false confessions; the recently discovered practice by whichcivilians were charged with the deaths of peoplewho were in fact killed by police officers; the high-profile police shooting of Laquan McDonald, which was almost certainly covered up by some officers at the scene. (And as I was writing this essay, several people coincidentally passedthis story along to me.) And these controversies all rest against the backdrop of the department's distressingly low clearance rate in homicide cases; police there have only managed to solve about 21 percent of the homicides this year. (For comparison, Washington, D.C, a city rarely held up as a paragon of police effectiveness, had a homicide clearance rate of 62% in 2015.) These are horrifying crimes that devastate communities, and too few people are held responsible. Or, as the late Harvard law professor William Stuntz wrote, "poor black neighborhoods see too little of the kinds of policing and criminal punishment that do the most good, and too much of the kinds that do the most harm."

Papachristos recently wrote about the consequences of this kind of bad policing in the New York Times. He cited a new study that found that a high-profile incident of police misconduct in Milwaukee — another deeply segregated city with a grim recent historyaround race and policing— made residents there far less likely to call 911 for help. That means that attempts at achieving justice often come via extra-legal means. "Research shows that urban neighborhoods with higher levels of [distrust in the legal system] also have higher rates of violent crime," he wrote. "When citizens lose faith in the police, they are more apt to take the law into their own hands."

Copyright 2016 NPR. To see more, visit http://www.npr.org/.