“I know a lot of people that’s been in jail. My dad, my uncle ... well, all my uncles. My cousins. But I’m my own person.” Fifteen-year-old Christel is one of several individuals featured in the Frontline documentary Prison State, offering a compelling personal account of incarceration’s typically anonymous collateral damage. The documentary centers on Beecher Terrace, a majority African American neighborhood in east Louisville. Kentucky pays roughly $15 million dollars a year to incarcerate the residents of Beecher Terrace and the immediately surrounding neighborhoods, according to Frontline.

Research on the social and public-health consequences of incarceration usually follows one of two arcs: tracing the adverse mental and physical outcomes for people who have been incarcerated, or else noting “spillover” effects that may include depressive or anxiety symptoms among the partners and children of those behind bars. Michelle Alexander, a professor of law at the Ohio State University and the author of The New Jim Crow, says kids living in communities like Beecher Terrace are “likely to attend schools with zero-tolerance policies, where police officers patrol the hall, where disputes with teachers are treated as criminal infractions, where a schoolyard fight results in [their] first arrest …[They] find that even at a very young age, even the smallest infractions are treated as criminal.” Indeed, in the documentary, after Christel begins skipping class, a school counselor tells Frontline, “She is so angry, Christel could be headed towards prison if she doesn’t get it right, right now.”

Residents of neighborhoods with high incarceration rates endure disproportionate stress, since these communities face disrupted social and family networks alongside elevated rates of crime and infectious diseases.  However, only recently have researchers considered how incarceration rate might serve as an ecological variable, rather than just as a something that shapes individual behavior. A new article in the American Journal of Public Health evokes environmental-health research by treating incarceration as a toxin. A team led by Mark Hatzenbuehler of Columbia University’s Mailman School of Public Health examines how higher than average “exposure” might impact community well-being in less direct ways.

To this end, the researchers trace mental-health outcomes of people who live in areas from which the incarcerated are heavily drawn, but who have not otherwise had contact with the criminal-justice system. Using a representative adult population sample from the Detroit Neighborhood Health Study (DHNS), the team organized corresponding prison-admission data from the Justice Atlas of Sentencing and Corrections for 28 zip codes. The prison admission rates across these neighborhoods varied dramatically, from 0.89 to 7.59 per 1,000 adults. The top quartile, containing neighborhoods with prison-admission rates above 5.12 per 1000 adults, was designated as the “high-exposure” group.

The team conducted in-depth phone interviews with roughly 1,500 DNHS participants in four waves spaced one year apart. Using criteria from the DSM-IV, researchers assessed generalized anxiety disorder (GAD) and major depressive disorder (MDD) symptoms over participants’ lifetimes as well as during the 30 days prior to each wave of interviews. Later, a blind clinical reappraisal study using a small probability sample confirmed the accuracy of their methods.

Even after controlling for individual- and neighborhood-level factors including history of incarceration, age, gender, race, personal income, trauma exposure, percentage of household incomes less than $25,000, and violent-crime rate, individuals living in high-exposure areas were significantly more likely to meet the criteria for MDD and GAD than were their counterparts in low-exposure areas. The team also tested neighborhood parole rates as another measure of incarceration, and found similar results to those achieved using prison admission rates.  They concluded that some elements of community supervision (potentially including the visibility of parole officers, the tone of their relationships with parolees and probationers, and the likelihood of inordinate police scrutiny due to previous arrest patterns in the area) may also be damaging to residents’ mental health. Finally, the results of their sensitivity analyses suggest that “the effect of neighborhood-level incarceration on mental health is similar for individuals with and without a history of incarceration.” These findings offer compelling evidence for a relationship between neighborhood incarceration and mental health independent of an individual’s experience with the prison system, indicating that the long arm of corrections reaches even further than previously thought.

Hatzenbuehler’s work is the latest in a growing trend of applying methods from the health sciences to complex social problems. Several public-health researchers have reasoned that contagion is an appropriate analogy for incarceration, due to the well-documented finding that proximity to an incarcerated (or “infected”) individual greatly increases the likelihood of incarceration. Ernest Drucker’s landmark A Plague of Prisons was the first mainstream work to apply epidemiological modeling to the problem of mass incarceration in order to trace its path through populations. Drucker created a profile for the disease, pointing out that young minority men are the most susceptible, making up more than 30 percent of the cases; poor urban neighborhoods are the hardest hit, with some communities having an infection rate of up to 90 percent; afflicted individuals are often socially marginalized and incapacitated for life, unable to find stable employment or housing; and the children growing up in affected families have shortened life expectancies and are six to seven times more likely to be infected than are children in unaffected families.

Others continue to strengthen the case for why incarceration is in many ways similar to a plague. Last fall, researcher Kristian Lum and others from the Virginia Bioinformatics Institute used a susceptible-infectious-susceptible (SIS) model of infectious-disease propagation to explain why even moderate differences in sentence length can cause the vast racial disparities observed in the aggregate prison population. An SIS model is nonlinear: The number of infected individuals and the number of transmissions are not proportional, due to feedback loops present as people transition between being infectious (incarcerated) and susceptible (out of prison, but at risk of going back).

The researchers in Lum’s study began by considering a puzzle in the data: The per capita rate of incarceration almost quadrupled between 1978 and 2011, and mostly affected African Americans. However, the authors note, “the increase in imprisonment of black males since 1980 was not matched by a similar increase in black-male criminality.” Their model considered incarcerated individuals “infectious” to those who are most affected by their absence, and adjusted the “transmission” probability according to relationship type and personal characteristics (children and male associates being the most susceptible). When individuals are released from prison, they return to a “susceptible” state, while their newly incarcerated friends and relatives introduce positive feedback into the system by becoming the source of additional infections. These feedback loops are responsible for the explosive transmission rate that continues to amplify racial disparities.

According to the Bureau of Justice Statistics, whites have a mean sentence of 14 months for drug possession, while blacks have a mean sentence of 17 months for the same offense—extending the infectious window by three crucial months per individual. Despite these enduring differences, Columbia’s New York Psychiatric Institute used large national data sets to demonstrate the lower rates of drug offenses and use for blacks versus whites in the United States. Applying epidemiological tools to the problem adds urgency to the argument for reform. The implications are clear, Lum and her fellow researchers write:

If incarceration risk is indeed propagated through social networks, our results predict that incarceration is self-perpetuating and changes to sentencing policy may have long-term unanticipated consequences. Indeed, harsher sentencing may hinder progress towards the intended goal of decreasing crime, creating safer communities and maximizing justice to the state, victim, and offender. Our model suggests that increased sentencing for an individual has negative effects that spread through social networks to affect families and whole communities. As a consequence, increased sentence lengths may create criminals from individuals who otherwise would have avoided criminal behavior.

These patterns reveal new depths of the incarceration-poverty problem, with health effects that may carry over into new generations alongside the related economic burdens. The cumulative emotional stress resulting from high incarceration rates in a community may generate feedback loops of its own—perhaps by inducing substance abuse and other behaviors associated with depression and anxiety, and in turn predicting further incarceration.