Michael was not in New York on September 11, 2001. But for years afterwards, when an elevator opened at work, he would imagine people on fire rushing out, their screams filling the lobby. When he closed his eyes, he would sometimes see limbs trapped in rubble, unattended by their bodies. He was plagued by moments of violence and destruction that he had not witnessed. On sleepless nights, he would wander the streets of his neighborhood, trying to exorcise other people’s demons.

Michael, who asked that his last name be withheld for privacy, is a clinical psychologist who works in lower Manhattan. In the years after the World Trade Center attacks, he treated hundreds of patients with acute and post-traumatic stress disorder. But it took him a while to notice that while the mental health of his patients largely improved with each passing therapy session, his own was deteriorating. By 2004—jittery, depressed, and unable to sleep most nights—he began to suffer panic attacks for the first time in his life. Increasingly, he had to withdraw from social events and public spaces. He asked a colleague to prescribe sleep aids and antidepressants.

Exposed each day to the distress of others, our second-line responders to tragedy—humanitarian workers, therapists, social workers, lawyers, and journalists—can develop traumatic stress disorders that mimic the PTSD of their clients, patients, and sources, down to the images of violence that can haunt a traumatized mind. Thirty to 40 percent of mental-health clients in the U.S. show symptoms of post-traumatic stress. By one estimate, as many as half of the psychotherapists who treat these patients could have symptoms of “secondary” or acquired trauma.

Currently, the best treatments for trauma require sharing the story of what happened. Talk is therapy—but when the things we share are horrifying, our listeners can be altered for the worse. In this way, individual trauma can morph into something collective.

* * *

Michael, who used to be a jazz percussionist, has the slight hunch of a former drummer and the soft white hair of a television physician. He tends to bob his head when he’s thinking. No clients attended his first post-9/11 clinic, he told me, because he’d set it up on the New Jersey waterfront in a makeshift hospital on the very day of the attack. Unable to reach patients that day, he could only watch as the familiar Manhattan skyline disintegrated.

The lull would not last. In the days following the attacks, Michael began seeing eight to 10 patients a day for roughly hour-long sessions, five days a week. All had acute stress disorder, the label given to the extreme emotional shock that follows a tragic event. If acute stress disorder goes unabated, it becomes PTSD. Some of his patients had escaped the collapsing buildings, or were first responders who sorted through the rubble after they fell. Many were plagued by images they could not forget, and by survivor’s guilt.

Characterized by volatility of emotions, hyperarousal, pervasive fear, and anxiety, trauma is an adaptive response whose aim is simply to keep a person away from similarly dangerous situations in the future. But in the long run, it can leave the victim emotionally distraught, forever alert to new, illusory threats.

The PTSD that follows is a disorder of association. Sounds, smells, images, and thoughts of the traumatic event will elicit a fight-or-flight emotional response long after the trauma has passed, as with the war veteran who jumps when a car backfires.

The best treatments for PTSD target these associations. Traumatized patients are encouraged to confront their associations, often by purposefully reliving the traumatic event in order to experience their full emotional and physical reactions. As patients recount their stories multiple times a session, week after week, the associations of the event can lose their force. Ideally, their reactions will weaken with each telling.

PTSD therapy can be a transformative process for the speaker, but its effect on the listener can be more complicated. “Service providers often must share the emotional burden of the trauma,” writes Brian Bride, a professor of social work at Georgia State University; they “bear witness to damaging and cruel past events, and acknowledge the existence of terrible and traumatic events in the world.” Hearing stories of suffering, in other words, can generate more suffering.

During lectures at psychiatric grand-rounds in medical centers, Michael often asks his audience of physicians and medical students to imagine a lemon. “Hold it in your mind,” he says, “See how yellow it is. Smell the citrus aroma. Now cut a slice off with a knife and take a bite. Taste the strong sour flavor.” When he asks people to raise their hand if they are salivating, nearly all do so. The point of the experiment is simple: What you think and imagine can result in a demonstrable, physical reaction. When a therapist for a patient with PTSD hears a story of violence, empathetic imagining can inadvertently trigger a physiological reaction similar to what the victim may have experienced: a racing heart, shaking hands, nausea, and other elements of the fight-or-flight response.

One of his patients, Michael recalled, was a construction worker who was tasked with clearing away rubble after the towers fell. One day, “he broke down crying in my office,” Michael said, because he had found a woman’s hand holding a child’s hand. “Just two hands together.” Michael went silent. “You know, right now, I’m feeling a response just telling you the story.”

Hearing stories of atrocity can also cause longer-term changes. Laurie Pearlman, one of the first psychologists to identify the phenomenon of secondary trauma, calls this effect “alterations to the cognitive schema.” Essentially, these stories can change the way the listener views the world, forcing him to recognize that his loved ones may not be as safe as he’d thought, and to face his own helplessness in preventing future tragedies.

This change in thinking can be gradual, and it can be a nearly unconscious process. When Pearlman began counseling victims of abuse in the 1980s, “I found myself affected by the trauma work in ways I didn’t expect or understand,” she told me. She started to view formerly benign settings as possibly threatening, and she found it hard to maintain her normally upbeat attitude. Today, she describes what happened to her as “disrupted world view” and “disrupted spirituality.” These symptoms of exposure to trauma stories, she believes, are the ones that can harm caregivers the most.

Whether an altered world view is ultimately destructive, Pearlman writes, “depends, in large part on the extent to which the therapist is able to engage in a parallel process to that of the victim client: the process of integrating and transforming these experiences of horror or violation.” She recommends that all trauma therapists undergo therapy of their own.

* * *

Trauma after tragedy is nothing new: Evidence of PTSD in soldiers and commanders is present in ancient Greek and Roman texts. The pages of many Greek tragedies, like Sophocles’s Ajax and Euripides’s Heracles, are rife with veterans maddened by war. In the Roman army, suicide attempts were, strangely, punishable by death—unless a soldier was found to suffer from shame, sadness, or “weariness of life.”

The notion that trauma can be communicable is much newer. Psychoanalysts since Freud have noted that therapy can result in “countertransference,” where a patient unearths, and becomes the object of, an analyst’s own neuroses, desires, or unresolved conflicts. But it was not until the Vietnam War that a greater danger was observed: that patients could, unknowingly, plant lasting images of violence in the minds of people who hadn’t experienced it.

“When the patient reports atrocities, where does the therapist begin?” asked Sarah Haley, a Veterans Administration psychiatrist, in a paper in the Archives of General Psychiatry in 1974. In the late years of the Vietnam War, Haley treated veterans with mental illness who described events of extreme violence and brutality, including cases where they were the perpetrators. She records the story of one patient, a Marine who led his squad in the destruction of a pacified village after booby traps were found in the surrounding jungle. “[We] blew them motherfuckers away,” he told Haley. The story, and others like it, left her “numb and frightened.” How do you treat such patients, Haley asked? “Perhaps,” she wrote, “we start by reminding ourselves that atrocities are as old as man.”

In 1981 Yael Danieli, a Manhattan-based clinical psychologist and former sergeant in the Israeli Defense Forces, published a review of therapist’s emotional reactions to working with Holocaust survivors and their children. The therapists, she wrote, often “found themselves sharing the nightmares of the survivors they were treating.” In nine months of treating survivors, one therapist reported having had only two dreams that were not related to his clients’ stories.

Another nearly fainted when a patient described seeing “children clinging to their parents’ bodies in mass graves.” Still another confessed that the first time he saw an identification number tattooed on a client’s forearm, he had “to leave in order to throw up.”

Danieli found that therapists, fearing their own reactions to traumatic content, dreaded meeting their survivor clients and often avoided discussing the Holocaust with them. Many expressed fears for their sanity. “I dread being drawn into a vortex of such blackness that I may never find clarity and may never recover,” a therapist told Danieli. “Once this little black box is open,” another said, “it’s worse than Pandora’s box.”

We now know that secondary trauma is a predictable consequence of working with traumatized populations. “Vicarious trauma is inevitable for people doing this kind of work,” says Jackie Burke, the clinical director of Rape and Domestic Violence Services Australia, a counseling service for victims of sexual abuse and family violence.

She believes that all members of her counseling staff suffer in some way from the trauma narratives of their clients. “But it will only rise to the level of dysfunction, of PTSD-like symptoms, for a few,” she says.

A review of case managers working in community mental-health services in the U.S. found nearly one in five had symptoms of PTSD. Similar rates of distress were found in mental-health workers who treated victims of the 1995 Oklahoma City bombing and survivors of Hurricane Katrina. One review of 100 sexual-abuse therapists found nearly half had secondary traumatic stress.

“What the research points out at this time is there is only one really reliable predictor of whether someone will get vicarious trauma,” says Burke. “And that is their level of exposure to trauma.” Because caregivers are exposed to traumatic content while on the job, “I conceptualize this first and foremost as a worker health and safety issue.” Burke’s organization employs two full-time “vicarious trauma” supervisors, who meet with each counselor on staff every week to monitor levels of secondary-trauma symptoms and, when symptoms appear to rise, to develop immediate treatment plans.

One of the most effective interventions, Burke says, is the “leave your work at work,” approach. “If you understand that the only thing that predicts traumatization is exposure to traumatic content, then the only way you can reduce risk is by reducing exposure,” she says. Organizations can do this by reducing the number of traumatized patients a worker sees, for example, or by enforcing limits on office hours. But she acknowledges that in many cases—including her own workplace—these measures aren’t always feasible. “For us it wasn’t possible to do that. Our online and telephone counseling services are 24-7.”

What they could try to control, though, was the exposure that was happening outside of work. When staffers told her that they kept thinking about their clients’ stories even when they were off the clock, Burke and her staff developed a 15-minute decompression exercise to help counselors “drop their worries about clients’ well-being with another worker who will be continuing on.” Before a counselor leaves her shift, she is required to tell someone still on duty all the disturbing things she heard that day, explain why they were disturbing, and list the things she’ll do to try to feel okay about it.

“This helps the counselor do two things,” Burke says. “First of all is to recognize that vicarious trauma is there and to think about what to do with that. And secondly, to feel like, ‘When I leave work I don’t have to think about all that stuff.’”

Burke’s team is now trying to retool the “leave it at work” process to work for counselors in more isolated settings, who may not have other staff to talk to after a difficult counseling session. “In that case, who do you leave your worries with?”

The answer for now seems to be: yourself. A new protocol asks counselors to record their feelings related to the work day before they punch out. The goal is immediate reflection to avoid repression of traumatic content and later rumination. So far, says Burke, “this seems to work just as well.”

* * *

Michael now admits that he likely had low-level PTSD for several years before his breakdown. “I felt my health eroding, but I could only diagnose it in hindsight,” he says. “I didn’t realize the full extent of what was happening.”

This will be the experience of most caregivers suffering from secondary trauma, unless they and their employers know to take the problem seriously. “You keep stuffing it down,” Michael says, “and try to work harder.”

In the end, Michael had to temporarily shutter his practice while he worked to get his PTSD symptoms under control. He’s back to treating trauma survivors, though, and considers himself a better therapist as a result of his struggle. “No one really understands what you are going through,” he says. “Not unless they’ve experienced it.”