I’m sitting in a conference room with pale green furniture around a long oval conference table in downtown Newburyport. Kelly Dunne sits to my left at the head of the table, wearing a black skirt and flats, her hair blond-tipped. A stack of files sits in manila folders in front of her. Detective Wile, in khaki shorts and running shoes, sits across from Howie Adams, a police sergeant from Newburyport, just back from vacation. Wile has a deep end-of-summer tan. Just outside, the Merrimack River joins the Atlantic Ocean, and white sailboats bob in aquamarine water in a ubiquitous New England summer scene. Signs advertise whale-watching tours and trips to Plum Island. Newburyport itself, which was once a run-down, blue-collar mill town, has gentrified, and now boutiques, organic restaurants, and local galleries proliferate. Probation, parole, the Merrimack police department, batterer intervention programs, and the local hospital all have representatives sitting around the table with us. Dunne and her colleague, Kate Johnson, head the meeting.
I’ve come to Newburyport because it is one thing to understand, in the abstract, that communication can improve a system, but I wanted to see precisely how a High Risk Team could strategize on the ground to get victims out of danger. I agreed to not identify anyone from Dunne’s caseload, and not to quote any healthcare team members who are barred by HIPAA confidentiality. If a detail was part of a public record, say a police report, then I’d use that, but otherwise, the primary way I was able to include specific details from any one case was if the detail is present in multiple cases; for example, there is a case where an abuser threatened to break CDs and slit his wife’s throat with them. Seems incredibly specific, identifiable to a certain couple, but as it turns out, Dunne sees this particular threat frequently (though presumably less often with the advent of streaming … which I realize sounds a little ridiculous, but is also probably true. Spotify saves lives.). This is how we agreed to maintain the safety of victims while also allowing me to see firsthand how it all worked.
Cases are referred largely by Dunne’s team or through one of the police departments, and the team votes on which referrals wind up in their roster. (Cases can remain active for years.) About 10% of all domestic violence cases are high risk for them. They scrutinize each case for changes—a new pregnancy, an attempt to leave, an abuser going off probation or parole, the violation of a restraining order, the loss of a job, an incendiary Facebook posting. They look at case histories and patterns of offender behavior through the lens of Campbell’s risk indicators. The day before, sitting in Dunne’s office, she handed me a stack of several dozen police reports with the names redacted to give me a sense of what she deals with on a daily basis. I sat on a soft couch in an unused room at the crisis center, white noise machines whirring simultaneously all down the hallway (so clients have privacy when they meet advocates). It reminded me of a yoga studio, all muted colors and soft lighting, and felt like a shocking inverse to the horror of what I sat reading:
“It’s not clear how I got on the kitchen floor but the next thing I remember is [X] on top of me strangling me with both hands.” “[X] made threats to me in the past about killing me, putting me in the chest freezer, & then taking my body out onto his boat & chumming me into the ocean. He also stated that he could kill me and put my body in his septic tank.” “[X] held her repeatedly against the heating pipes.” “[He] would … gut her and hang her up like a deer to bleed.” “If I decide to bring home another woman with me, you’ll do what I tell you to do with her. I am your master, you are my slave. If you don’t do as I say and keep me happy, I will kill you.” “[X] frequently makes threats to [her] life, including breaking her CDs in the car … and threatening to slice her neck open with them while she is driving.” “[X] held her hostage with a rifle … [and] stated he would ‘set up’ over 1,000 yards away and ‘take her out.’ ”
Earlier, I’d spoken with one client whose abuser, her ex-husband, had been put on GPS monitoring by the team. (She spent more than a year in protection as part of the Commonwealth of Massachusetts Address Confidentiality Program.) She told me, “He doesn’t have the judgment of someone who has something to lose.” It reminded me of James Baldwin, The Fire Next Time. “The most dangerous creation of any society is that man who has nothing to lose.”
The day I attend, Dunne and Johnson have fourteen cases to discuss. One of the first problems they encounter involves the privacy of medical information. It was a case of severe strangulation by her ex-husband. The woman had called the crisis center, gotten an active restraining order against him, and was put in the high risk roster. He, in turn, was put on probation. But in the past week, he’d called her and threatened suicide. She phoned the paramedics and he was taken to the hospital. Afterward, he was jailed for violation of his probation, where he now was being held. The team has to strategize about what might happen upon his release. Can he be referred involuntarily to a psychiatric ward? Was there any indication of his demeanor at the hospital? The hospital representative, Moe Lord, is mostly silent even though she may well have been there when he came in and recognized his name from an earlier high risk meeting; HIPAA laws restrict her from saying almost anything. She tells Dunne she can’t speak to his demeanor. She says generally she often sees guys come into the hospital wearing GPS ankle bracelets and knows it likely means they’re abusers on probation. She also sees women with suspicious injuries, but unless a patient discusses the situation with her, she can’t even call the police on a restraining order violation. Still, she’s here because getting a sense of these cases enables her to try to intervene with known victims should one wind up in her emergency room. She’ll at least be able to share information about Dunne’s agency, and give a sense of how that victim can get specific help.
“Is there a mechanism at all for when [abusers will] be discharged, to call the police?” Dunne asks Lord.
“If we know about it and he gives us permission to speak,” Lord says.
“So he could walk right out of there?” Dunne asks, referring to the hospital. This is a dangerous scenario. He’s broken his restraining order, and the victim has called the police on him, which is likely to escalate the situation, and enrage him even more.
Lord nods.
Dunne’s face flusters with frustration. She glances down at the file in front of her, forehead creased in concentration. She has always struck me as uniquely unflappable, steady under stress.
A probation officer in attendance says the victim came to see him after she called police, and Detective Wile, it turns out, knows the abuser. He offers a history of accusations from other victims (which I am not allowed to include here). Wile has lived in this part of the country for his whole life. He seems to know every family from every small town around here and in meetings he will often wax on about this family or the other and their years of being in and out of trouble, or years of drug abuse, or what brother married some sister from a family two streets away and now he’s watching the kids of those kids start to get in trouble. He’s fond of the word “knucklehead.”
“I would think the exception would be the Tarasoff warning,” Dunne says, finally. She is referring to a warning required of mental health professionals when there is a credible danger to a victim. It’s sometimes called, simply, the “duty to warn.”
Lord considers this, nodding, and says, “If I know about it.” If she knows, in other words, that an abuser she recognizes from their caseload could be a credible danger to a victim upon his release. “I think we just figured a work-around,” Dunne says to Lord. “If in the future we had a more dangerous case, if you heard from probation there’s just a warrant, you can’t do anything. [But] if you heard the information a little differently … like he’s made a threat against her life, that’s a Tarasoff.”
Lord nods at the possibility.
The team strategizes a coordinated plan for each case. Sometimes, the police do extra drive-bys or home visits. They’ll note the cars at the house, lights on or off that are unusual. I remember one story of an officer who saw a light on in the upstairs window of a home he was tasked with watching. He stopped in at the house, asked the victim if everything was okay. It turned out her kid had turned on the attic light. When the officer drove away, he circled back—it’s a common stalking tactic for abusers to be watching for a police drive-by and then show up just as the officer rounds the corner. But the police in this part of the country are well aware of this tactic from the High Risk Team. So the officer drove around the block, circled back to the house two minutes after he left, and found the abuser there, just getting out of his car.
Other times, an abuser might be put on GPS, or have exclusion zones—usually entire towns—he is barred from entering. They might place a victim in transitional housing or help with legal fees and representation or give safety training. They might change locks or provide new cell phones to the children and the victim. “If you care about the long-term health of a victim, not having them killed is not enough,” Dunne says. “When that offender goes to jail, her physical safety may be okay, but her life might unravel [with] the loss of support. You have to restore that victim back to the state they were in before that violence occurred.”
To Dunne, this is critical. Victims often come with their own set of issues. Addiction, poverty, unemployment. Dunne is not trying to fix every aspect of someone’s life. She’s trying to get them out of danger to a space where they may be able to think through solutions to more systemic problems—like employment or addiction. To maybe give them the emotional and physical and mental space to address some of these other issues.
“The key with domestic violence,” says Dunne, “is addressing it in misdemeanor phase.” One of the more challenging elements of domestic violence is that ideally you want to stop abuse from escalating. But for that to happen, misdemeanors need to be taken far more seriously than they generally are in intimate partner violence cases. Like the case in San Francisco with Tari Ramirez and Claire Joyce Tempongko. Some of the most extreme violence in cases like this are misdemeanors, for which abusers are given shockingly light sentences. Ramirez, recall, was given only six months. Even Donte Lewis was given only four years for kidnapping his girlfriend and hitting her so hard on the head she foamed at the mouth and passed out. And many perpetrators like Ramirez go straight from a misdemeanor to a murder. But for the judiciary, the challenge is what to charge an offender with and how much the court can and will respond in an attempt to stop the behavior.
One of the most effective tools for the High Risk Team is a Massachusetts bail statute called a 58A, or dangerousness hearing. A standard bail hearing is meant to determine an offender’s flight risk, whereas a 58A can be requested by the district attorney and allows defendants, even those with clean records, to be held in the misdemeanor phase without bail until trial if they are deemed a sufficient threat to an individual or community. The statute might have prevented the release of William Cotter and saved Dorothy’s life, but it was seldom used for domestic violence cases at the time and so not on Dunne’s radar, and the details of the Cotters’ situation were scattered among agencies. In Massachusetts, offenders can be held for 180 days. “A lot of violence happens between arraignment and disposition,” Dunne says. “We contain the offender so the victim doesn’t have to be contained.” Few states have such a clear dangerousness statute, and in her trainings now Dunne encourages advocates to search their states’ bail statutes for something similar. Many advocates simply don’t know to look, though even when they do go back and research their own states’ bail statutes, they’re almost sure to be disappointed. In April 2018, Pennsylvania became only the second state to pass such a statute allowing a judge to take into consideration, specifically, the dangerousness of a domestic violence abuser.1
The 58A belongs to a category of bail statutes known as preventive detention, and Dunne says in her trainings across the country, she almost never meets other advocates who have something like the 58A. Most, in fact, ask her how they can get such a statute passed in their jurisdictions. “Many states have preventive detention statutes,” said Cherise Fanno Burdeen, the chief executive officer of the Pretrial Justice Institute, an advocacy group that works with communities on effective bail practices. “But they are sorely underused. The systems use work-arounds which unfortunately don’t always work, meaning dangerous people leave jails every day, unsupervised.”
Preventive detention statutes emerged from federal legislation called the Bail Reform Act of 1984, which allows a defendant to be held pretrial if he or she is deemed dangerous enough to another person or to a community. A determination of dangerousness includes factors like the nature of the crime, the evidence against the defendant, and the history of criminal activity, among others; most often, these statutes have been used in gang or drug cases, though Massachusetts has seen a marked increase in their use for domestic violence.
While no one tracks how often such hearings are held—either in Massachusetts or nationally—each state differs slightly in how preventive detention statutes are implemented and on what grounds. One common thread, however, is the controversy of whether to use preventive detention at all. “The Constitution tends to frown upon punishing prospective behavior,” Ronald S. Sullivan Jr., the director of the Harvard Criminal Justice Institute, told me. “We punish past behavior that’s been proven. Here we’re keeping people in jail because we think they’ll be dangerous.” But Viktoria Kristiansson, an attorney adviser for AEquitas, cited the importance of the dangerousness hearing, claiming it “automatically provides a different context for a judge to analyze the evidence.”
Holding an abuser before trial has helped to keep victims out of shelter; it provides them with the time to find alternative housing, save some money, find counseling, and perhaps find a job. “We know that arrest, in and of itself, is protective,” Dunne told me. “You’re trying to disrupt that escalating cycle of violence.” Holding an abuser in jail prior to a judgment in court allows victims time to regroup, to put their lives in order prior to the trial when the danger levels increase again. Dunne says it’s been critical to their success. And of the offenders monitored by GPS bracelets in Dunne’s caseload, not a single one has re-assaulted and nearly 60% of them have been held pretrial using a dangerousness hearing in her jurisdiction. Though no one has tracked how much the 58A was used before the High Risk Team was created, anecdotally Dunne said she saw “maybe five in three years” before Dorothy’s homicide. “Now we see two a month.”
The last case of the day is an immigrant woman and her abusive partner. He is currently in jail for assault and battery—which jeopardizes his immigration status—but the couple has a small child currently staying outside the country with the abuser’s family. The family has threatened that if the charges against the abuser are not dropped, the young mother may never see her child again, and if her partner is deported before she gets her child back, she also may never see the child again. What this means is that she cannot testify without potentially losing her child; she cannot, in fact, appear to be siding with the prosecution at all. It’s the kind of insurmountable situation that both victims and prosecutors alike find themselves in. Though it hasn’t happened yet, the team fully expects her to recant. Sergeant Adams describes the couple’s history in the affidavit from the night of the arrest, most of which I cannot recount here, except to say that he had so isolated her that he wouldn’t allow her to make calls on her cell phone to anyone but him, and he had set up cameras inside their home to keep tabs on her.
Wile suggests an amendment to the complaint that will add several additional charges that, he says, “will give the DA’s office something to work with. If you can get eight or nine charges out of one incident, it’s a better shot that she won’t have to take the stand, and he’ll plead on some of those.” Wile is referring to several different elements here. The first is to try to charge an abuser with as many allegations as possible in order to negotiate some kind of plea deal, even beyond the violent incident. Are there possible drug charges? Is there an illegal weapon or firearm in the house? This also offers a better chance that at least some of the charges might stick. The other element Wile is referencing is called evidence-based prosecution. Meaning “evidence” rather than “witness” based. A prosecutor can offer up enough evidence in court that a witness wouldn’t have to come and testify in front of her abuser. Such evidence may include photos, affidavits, witness testimonies, prior records, or 911 tapes.
This is how, for example, Stacy Tenney and Michelle’s family could have prosecuted Rocky Mosure with or without Michelle’s testimony had they ever found that snake. They could have coupled that with an affidavit about the time Rocky threatened them all with Michelle’s grandfather’s gun, or, had they known, about the many times he’d stalked Michelle on her way to and from school or kidnapped the children as leverage. Had they been able to hold him for dangerousness prior to trial, they’d have learned he was in and out of work, on and off drugs. They could have known all of this, acted on this and perhaps more if they’d had, back in 2001, a High Risk Team sharing information, creating a more complete picture of the situation across bureaucratic borders.
Evidence-based prosecution for domestic violence cases existed back when Michelle recanted, too. Ellen Pence, the advocate in Duluth, Minnesota, who’d created the Power and Control Wheel, was advocating for it way back in the 1980s. But it wasn’t until Casey Gwinn, then a prosecutor in San Diego, took notice of her efforts and began to bring one such case after another to trial in his jurisdiction that the evidence-based movement in domestic violence really took off. Gwinn traveled to Duluth and met with Pence, learned about her advocacy, and returned to San Diego just in time to try his first evidence-based domestic violence case—against a sitting judge named Joe Davis. Davis’s girlfriend had recanted, and then disappeared. But Gwinn pressed charges anyway.
And in front of local media and television cameras, he lost the case.
It was a humiliating defeat. The whole trial had been intensely followed in the region, given that it was against a judge, and Gwinn told me he just “made a fool of himself … I didn’t know what I was doing.”
But after the Davis trial, the San Diego city attorney at the time, John Witt, called Gwinn in to his office and told him it was going to be difficult for both of them, for their entire office, for a while, but that he believed in what Gwinn had attempted to do. “He told me to go out and figure out how to win these cases,” Gwinn said.
Gwinn began ordering 911 tapes in all domestic violence cases—something he’d never done prior to Davis. And he asked the police to take pictures of everything: the crime scene, the victims, even the perpetrators if they were raging in the backs of police vehicles. Any possible shred of evidence that existed, Gwinn wanted. He began to go out to roll calls at local police departments, asking them to gather more and more evidence. When one sergeant told him his entire mission was bullshit, that Gwinn was never going to prosecute these cases, Gwinn created a messaging system to let police know how their cases were resolved.
Gwinn tried twenty-one cases in a row, all domestic violence misdemeanors. All without the victim testifying.
He won seventeen of them.
By the time VAWA passed in 1994, Gwinn had trained attorneys around the country on evidence-based prosecutions in domestic violence cases. (It’s a bit of a misnomer to call them evidence-based since all trials are, technically, based on evidence.) He came to fervently believe that if we could prosecute murderers without a victim’s cooperation, we could prosecute batterers. In 1996, Gwinn was elected city attorney of San Diego and he made good on a campaign promise to devote 10% of his entire office to a domestic violence unit. By now, jurisdictions around the country came to him for training. Gwinn said they went from prosecuting fewer than 5% of domestic violence cases in the 1980s to some jurisdictions prosecuting 80% of such cases by the late 1990s.2
Then, in 2004, Crawford came.
In Crawford v. Washington the Supreme Court ruled that cross-examination is required of witnesses at trial unless a witness was unavailable (e.g., sick or dead). The court said that a defendant had the Constitutional right to face his accusers, that testimonial statements by witnesses who did not appear at trial were hearsay. And hearsay was not admissible.3 This meant victims who were too terrified to appear in court but were otherwise healthy could no longer allow prosecutors to use their statements.
Post-Crawford, there is still some room for state courts to determine admissible evidence using their own discretion, but generally speaking Crawford had a profound effect on the movement of evidence-based domestic violence cases across the country. These days, victim statements are often inadmissible in court proceedings if a witness is uncooperative (as happens in as many as 70% of cases).4 “The barrier to evidence-based prosecution is not about evidence,” Gwinn said. “It’s not about the viability of winning these cases. It’s about cultural norms and values. And at the heart of it is a stunning amount of misogyny.”
One of the critiques levied at the Newburyport model is that scaling up to a busy urban setting where resources are scarce and domestic violence calls nonstop would prove difficult. Mark Gagnon, the former chief of police in Amesbury, dismisses this criticism. “A bigger community’s going to have more resources,” he says. “It can be done, just at a different level.” Dunne recognizes the challenge of scaling up, but says it’s a matter of dividing territory into manageable jurisdictions. Urban areas in Massachusetts like Framingham, Lynn, and Cambridge have undergone trainings by Dunne and Wile and developed High Risk Teams. “I think one of the benefits of the model is that it not only changes things for the victims, but it changes things for everyone,” said Mary Gianakis, former director of Framingham’s Voices Against Violence and head of their High Risk Team. “It changes the way every spoke on the wheel approaches domestic violence … I also think it sends a clear message to perpetrators. That as a community, we’re not going to tolerate this kind of violence … It’s an important message, because it changes the culture.”
Today, Kelly Dunne and Robert Wile have trained tens of thousands of people across the entire country. And groups from California to Louisiana, Florida to Illinois, have asked for the training. Campbell did the research, but Dubus and Dunne put theory into practice, she said. “They were informed by my work, and now I’m informed by theirs,” Campbell told me. Former vice president Joseph Biden championed the Amesbury program in October of 2010, when he honored Suzanne Dubus at a White House event to mark Domestic Violence Awareness Month. “We need to replace what we have been doing and replicate this kind of success,” he said to those gathered.
To Dunne, their success demonstrates how a seemingly intractable problem can be addressed at relatively low cost, through coordination, the sharing of information and pure vigilance. “With Dorothy’s case we were looking for a fire alarm while we were in the fire,” Dunne said. “And you can’t do that. You need a system already in place.” Prior to the team’s formation in 2005, the town of Amesbury alone had on average one domestic violence homicide per year. Since they began, Dubus and Dunne have not had a single homicide in their caseload. What is of equal importance to Dunne, though, is that they have had to put fewer than 10% of the survivors in shelter; before 2005 that number would have been above 90%. To Dubus, it seems obvious to create a model in which the victims are protected rather than banished. “Here’s the outrage,” she told me. “It’s really cheap to do what we’re doing. It’s a lot cheaper than murder investigations and prosecutions and jail time.”
In late 2012, the Department of Justice’s Office on Violence Against Women in Washington, D.C., earmarked half a million dollars to replicate the High Risk Program, along with another out of Maryland called the Lethality Assessment Program. Twelve sites around the country were originally considered, including Rutland, Vermont; Brooklyn, New York; and Miami, Florida. In the end, only one test site got the green light from OVW to replicate Dunne’s program: Cleveland, Ohio.5
In the state of Ohio, between July 1, 2016 and June 30, 2017, there were 115 domestic violence homicides.6 The year 2016 alone had more than seventy thousand domestic violence incidents (just over half had charges filed).7 “Seventy thousand calls,” said Tim Boehnlein, of the Cuyahoga County Victim Witness Service Center, who is also one of the two leaders of the High Risk Team in Cleveland. (The other is the Domestic Violence and Child Advocacy Center.) “That’s a lot of people whose lives are being affected by domestic violence, a lot of families.”
The city of Cleveland is divided into five districts within the police department; the three inner districts had domestic violence officers, but neither the first nor the fifth districts—the city’s far east and far west sides—had the capacity to deal with the overwhelming number of domestic violence calls daily. So that’s where the city targeted its new high risk experiment, beginning in October 2016.
Dunne called me at home one afternoon just after she’d returned from one of her training sessions in Cleveland. “There’s someone you have to go meet,” she told me. “Don’t ask me anything. Just go.”