Tens of thousands of inmates in the United States still spend months and even years in solitary confinement, despite mounting evidence that the practice has little impact on the long-term safety of prisons but detrimental and irreversible effects on the health of the person subjected to the punishment.

The latest study to quantify the widespread use of confinement was released Wednesday, and comes from Yale Law School and the Association of State Correctional Administrators. In the study, titled “Aiming to Reduce Time-In-Cell,” the authors analyzed survey responses about its use from 48 jurisdictions—the Federal Bureau of Prisons, 45 states, the District of Columbia, and the Virgin Islands—which house about 96 percent of state and federal prisoners. The responses corresponded to its use during the fall of 2015.

Although the survey’s breadth is significant, it’s worth noting that it still doesn’t give the full picture of the solitary confinement in the United States—a practice that some corrections officials are working to end. It excludes local jails, which hold hundreds of thousands of people, as well as juvenile, military, and immigration-detention facilities. Additionally, not all respondents provided complete answers, so some calculations are restricted to locations for which a particular set of data was available during the survey collection period. Its findings, then, must be read with a caveat: that there are many more Americans who’ve experienced solitary confinement than the ones documented in the study.

According to the total survey results, 67,442 prisoners were “held in prison cells for 22 hours or more for 15 continuous days or more,” which is the working definition of solitary confinement used in the survey and in the study. In all, detailed information on length of stay, gender, race, age, and mental illness were provided by 41 jurisdictions, for a complete sample population of over 54,000 inmates who spent time in solitary—often referred to as restricted housing, administrative segregation, or “the box.” Of those, 29 percent spent one to three months in solitary confinement. At the other extreme, 11 percent endured three continuous years or more in isolation. Among all respondents, the share of prisoners in solitary in a given facility ranged from under 1 percent to more than 28 percent of the general population.

Because of differing survey methodologies, the study’s authors cannot directly compare the total fall 2015 findings with those from 2014, the last data-collection and analysis period for an earlier version of this study. But there was a decrease in use in some states: Utah, Colorado, and South Carolina.

For many corrections officials, though, decreasing the number of inmates held in solitary isn’t enough. Around the country, they are largely in agreement that the practice must be reformed and in some instances phased out. In 2013, ASCA enacted new guidelines regarding solitary that provide advice on acceptable accommodations, lengths of stay, and restrictions related to inmates with mental-health issues. And earlier this year, the American Correctional Association approved new standards that limit its use significantly.

One of the most vocal reform proponents is Rick Raemisch, the executive director of the Colorado Department of Corrections, whose state includes the U.S. Penitentiary Administrative Maximum Facility in Florence, a supermax prison. He has used his state as a testing ground for many of the reforms being considered across the country. He’s also become convinced that the future of solitary confinement is one in which the practice is not only rare, but is accompanied by support services for inmates and guards alike. We spoke earlier this week. Our conversation has been edited for clarity and length.

Juleyka Lantigua-Williams: Why are you such an advocate for significantly reducing or eliminating the use of solitary confinement?

Rick Raemisch: If you get right down to basics, what we had talked about here was that our purpose, our mission, is public safety. With 97 percent of the inmates returning to their [home] community, it’s just important that we send them back better than when they came in. When we took a look at the use, or overuse, of solitary confinement, we felt that the reason it was being so overused was because it was helping a facility run more efficiently. That’s a noble goal, but it’s not our mission. Our mission is public safety.

Lantigua-Williams: How did putting some inmates in solitary confinement help run a facility more efficiently?

Raemisch: You have to go back historically on how these supermaxes got built to begin with, and how, at least I believe, solitary started to become overused. It was the get-tough-on-crime times. Our prisons were full across the nation. States were losing control of their prisons, and so they bought into the fact that in order to have more control we need to segregate [those who were] at the time considered the most violent inmates, which was originally the true purpose [of solitary confinement]. But that morphed into [using it on] those that were disrupting the facility, those that were involved in smuggling contraband. Things of that nature, violating policies.

It became much more than it should have been, and that’s when we started looking at individual cases. Almost all states, including Colorado, got rid of [using confinement as punishment for acting out], and almost all states have the policy that solitary confinement is only to be used [selectively] for the most violent inmates or those that are involved in the disruption of the day-to-day operations of the institution.

Well, that’s all my inmates. That’s what they do. The point is that you can find any reason to put someone into solitary if you want. I have called it the “steel-door solution”: If you have someone that doesn’t follow the rules, you throw them in a segregation cell, slam the door shut, and walk away. It used to be the standard line was, “Well, there’s no data that shows that it’s harmful.” Well, now there’s all kinds of data. It’s not just mental—it’s physical problems.

Does solitary work? No. It works for one purpose, really, which is if you have a very serious incident occur, you need to put that person somewhere until you can figure out what happened and to start to address the cause. What we have found is that our data has shown that the less you use it, the safer your facilities are, and that the safer your facilities are, the safer your community is once they get out. We’ve tried to build around positive reinforcement versus solitary confinement, which by any means just isn’t effective.

Lantigua-Williams: You mentioned the importance of not having these individuals go directly from solitary confinement back into their communities. What are the significant differences when that happens versus when someone comes out from a prison’s general population?

Raemisch: When I first got here, I heard stories of individuals being taken out of solitary in chains, then put into street clothes, [then] put back into chains. We’re talking leg chains, belly chains, hand restraints. Doing the inmate shuffle, as I call it, because you can’t walk correctly when your legs are shackled. Putting them on a public bus, taking the shackles off, and getting off the bus.

I told people: “If I was the bus driver, I’d stand up and scream at the top of my lungs, ‘Run!’” When we talked internally about it, the cynical ex-cop in me said, “Let me get this straight. You’ve deemed someone too dangerous to be in the general population of a maximum-security facility, being supervised by those that are trained to supervise them, yet you release them directly into the community? What the hell’s the matter with us?” You might as well just stamp “Return to prison” on their forehead.

At what time did it become okay to lock someone in a seven-by-13-foot cell or smaller for 23 hours a day, seven days a week, for over 20 years? At what time did it become okay to take someone that’s mentally ill and put him in a seven-by-13-foot cell for 23 hours a day letting the demons chase them around in that cell? At what time did it become okay to take someone that has spent years in solitary confinement and release them directly into the community?

We’ve changed all that. We haven’t released anybody from segregation—restrictive housing, as it’s called now—directly from those units into the community from the state prison system since March of 2014.

Working with the legislature, our governor [John Hickenlooper], and the ACLU, we were able to pass a statute that, for the most part, prohibits putting someone who has been classified as seriously mentally ill into solitary. We’re the only state in the nation that, when someone goes into solitary, they know when they’re coming out. That’s a maximum of a year, and even then, that year is for the most serious offenses such as homicide or serious assault. But even that we’ve taken a step further. We put [developmental] program tables outside of our restrictive housing units, where on a volunteer basis, if inmates want to come out for programming—not just for that one hour [out of the cell] per day [they are designated], but to come out for additional programming plus the one hour per day—they can do that. We’re getting away, for the most part, from a legal definition of what’s considered restrictive housing or solitary confinement.

Lantigua-Williams: Are there any options for therapeutic services offered, now that you’re working based on the recognition that solitary can actually be more damaging?

Raemisch: Yes. In fact, we have two institutions that are dedicated to those that have mental-health issues. Solitary has been banned at both of those facilities. We also have policies that if a person is involved in a disciplinary incident, a team looks at the incident, and if it was determined that a mental-health problem caused that incident, they’re taken out of the disciplinary process and put into treatment programs. We started what’s called residential treatment programs, which are becoming a national model.

I tell people that when we started these reforms not only did we not have a map, we had no road—because none of this has been done before, so we had to invent things based on the experience of our clinicians. Some things would work. Some wouldn’t. But we developed what’s called a 10-in-10 program, which is starting out of their cells for 10 hours of therapy per week and then 10 hours of extracurricular activities. The purpose, of course, is to get them back into general population. We also had to develop step-down programs for those that had been in solitary for periods of time. You just can’t simply open the door and say, “You’re going back to general population. Have a good day.” That’s not going to work.

The first step-down program is really just more of a socialization period when they’re out with other inmates a minimum of four hours per day. Once the clinicians feel that they’re ready to move on, they go into another step-down program where there’s actual therapy programming: anti-anger, cognitive programming, things of that nature. We looked at it systemwide and saw the [violent] incidences were down.

The next question is: If you don’t use solitary, what do you use? What we developed were what we called de-escalation rooms in these facilities, and they’ve been so successful we’re going to put them in probably all our facilities. They are basically time-out rooms [that inmates can request to use]. There’s music being piped in, and there’s stress-relieving instruments in there such as stress balls, a chalkboard if they want to write their thoughts out. They’ve been very successful.

Lantigua-Williams: How do you see solitary confinement trending nationally? Do you think that there’s a widespread-enough recognition that there needs to be considerable change?

Raemisch: There is. Nationally, there’s a movement now to change what we’ve done. The data, I think, is so powerful that it’s kind of, “What the hell are we doing?” At least that’s the way I look at it. I think that, again, when you get back to “At what time did this become okay?”—the answer is it never should’ve been. A person’s going to be coming back from prison one of two ways: either prepared and ready to be a good citizen to the community, or they’re going to come back angry. If they come back angry, we’re just going to have more victims, so my sole purpose in doing this is to try not to have more victims. It’s safer for my staff and it’s safer for the community, and that’s my job. That’s what we’re supposed to be doing. That’s what our mission is.

Lantigua-Williams: Do you think that this type of reform could inspire a statewide decarceration movement?

Raemisch: In all states, including Colorado, corrections is the largest mental-health institution in each state. There’s no question about it. Over 30 percent of my inmate population has a mental-health issue. Over 10 percent of that 30 percent are what would be considered seriously mentally ill. I think what’s happening is that if we’re sending back better citizens than before, it just automatically means we’re going to have less coming in the front door again. Our recidivism rate is going to be lower.

The other is that there just seems to be a movement right now. It’s a pretty exciting time to be in the criminal-justice system because that door’s opened a bit—about finally realizing that this mass incarceration is just not working. Enough is enough. We’ve got to change the way we’re doing business.