How prisons keep their suicidal inmates on watch

Suicide watch can vary across prisons and jails, Tartaro said, but corrections facilities are increasingly moving to a new way of doing things.

Traditionally, inmates are locked in a bare cell by themselves, she said. Many are given paper gowns or tear-proof clothing to prevent them from creating a noose. Their belongings are confiscated.

"Suicide watch has wound up being an incredibly bleak environment," she said. "There's not going to be a thing in there to help you take your mind off of that. It doesn't bode well for you being able to get better."

But some correctional facilities are trying to move away from that model, Tartaro said, by allowing certain possessions -- like books -- and by no longer keeping these inmates in isolation.

At Tartaro's recommendation, in 2011 the Atlantic County Jail in New Jersey began pairing inmates they were concerned about with more stable ones, similar to the "buddy cells" (PDF) in some Australian prisons, she said.

The Pennsylvania Department of Corrections overhauled its mental health programs after a 2013 investigation by the Department of Justice, which examined how it treated mentally ill inmates and how it used solitary confinement. The new programs included a 75-hour training for inmates to become peer counselors, a paying job.

"You have an individual who comes to a prison mentally ill, not only gets in recovery but learns to help others," Pennsylvania Department of Corrections Secretary John Wetzel told CNN last year. "That's great stuff."

A small study by researchers at the Federal Bureau of Prisons found that trained inmate "observers" could reduce the amount of time suicidal prisoners remained on watch by nearly 45 hours on average. The study also highlighted a potential yearly cost savings of a program that reduced the amount of time prisoners stayed on watch: $300,000 in overtime pay alone.

"It's very difficult to kill oneself when somebody's right there," Tartaro said.

"That's an important lesson to learn from Aaron Hernandez. He was alone in that cell."

Suicide watch protocols can suffer wherever resources are low, Tartaro said, especially for many local jails that have transient populations and may rely on limited local government funding.

"Jails have special challenges," Tartaro said. Assigning a jail guard to watch one inmate may be a bigger drain on resources, she said, than in a prison where inmates are housed long-term.

Two-thirds of the jails surveyed in a 2010 report by the National Institute of Corrections did not regularly assign suicidal inmates to "suicide-resistant" rooms. Because a majority of inmate suicides are hangings, the report said, an ideal room is clear of "anchors" to which bedsheets and fabric can be tied.

About 8% of inmates in local jails were on suicide watch when they took their own lives, according to the 2010 report, which analyzed nearly 700 jail suicides in 2005 and 2006. While 93% of jails had a suicide watch protocol, fewer than 2% had the option for constant, around-the-clock observation; most facilities checked on inmates every 15 minutes.

In the report, experts recommend "close observation," checks every 10 minutes or less for inmates who have recently tried to hurt themselves or who have expressed thoughts of suicide. "Constant observation" is for those who are in more immediate danger, whether they are threatening to commit suicide or are already harming themselves.

But some say 10 minutes could be much too long. It only takes three minutes after hanging oneself to cause permanent brain damage and 5-7 minutes to die from it, according to a report (PDF) on prison suicide by the World Health Organization's Department of Mental Health.

Suicides accounted for 7% of deaths in state prisons in 2014, according to a report (PDF) published last December by the US Department of Justice's Bureau of Justice Statistics. This comes to about 20 per 100,000 state prisoners.

The suicide rate for local county jails is more than twice that, according to Lindsey Hayes, a suicide prevention expert with the National Center on Institutions and Alternatives. He previously consulted with the Massachusetts Department of Corrections, which runs the prison where Hernandez was being held.

Hayes said in an e-mail that he's confident the agency will conduct a thorough investigation into Hernandez's death, "which should include ... whether or not his death was preventable."

Tartaro said, beyond suicide watch, two important elements of suicide prevention are training for officers and the screening of prisoners. Her previous research found that jail officers received less training than police officers, though this was not specific for suicide training.

After the Pennsylvania Department of Corrections revamped its mental health programs, it took only a year to train all 15,000 to 16,000 staff members in Mental Health First Aid, Wetzel said.

Screening for suicidality can be difficult because prisoners, who are mostly men, may not want to admit there's a problem, Tartaro said; being put on suicide watch can damage their reputation. But she also said mental health screening can be targeted at certain times: around conflicts with other inmates, problems at home and especially court dates.

"We do know that the days before and after a court appearance are high-risk times for inmates," said Tartaro.

Hernandez was found not guilty on Friday on two additional counts of murder. His attorney, Jose Baez, said in a statement Wednesday that nothing Hernandez said to his family or legal team "would have indicated anything like this was possible."

But his acquittal may not have been seen as a complete win by Hernandez, who already had a life sentence without parole, Tartaro said. He may have lost something to fight for, she said, as well as the opportunity to see his wife and daughter in court.

"It was a very hollow victory for him," she said.

The recent inmate suicides may not show the bigger picture, according to Hayes.

While suicide rates within correctional facilities have risen over the past few years, Hayes said, the rate is still much lower than revealed by the studies he conducted in the 1980s.

"As such, although not all inmate suicides are preventable, many of them are," he said.