Skip navigation

Secret BOP Document Raises Risk Factors, Security Levels of Prisoners

A secret Federal Bureau of Prisons (“BOP”) document, obtained by ProPublica, is being used to evaluate the security levels of prisoners, leaving some who qualified for release to home confinement stuck in prison during the COVID-19 pandemic with little explanation of how they were evaluated.

The First Step Act of 2018 required the BOP to create a risk assessment tool and publicly post it on its website for public comment. The tool, dubbed “PATTERN” (Prisoner Assessment Tool Targeting Estimated Risk and Needs) by the BOP was posted on its website and public feedback raised concerns about racial bias and lack of transparency, BOP spokesman Justin Long acknowledged. But the feedback from the public did not call for categorizing prisoners under a harsher standard.

The secret 20-page document adopted by the BOP in early 2020 includes a method that sets a security level for all prisoners. Factors like age, classes completed in prison, and any history of violence increase or decrease a prisoner’s score, classifying them as either minimum, low, medium, or high risk.

The previously unpublished document does not appear to be finalized by the BOP and its existence surprised prison reform advocates and lawyer. “It really tanks the whole enterprise if, once an instrument is selected, it can be strategically altered to make sure low-risk people don’t get released,” Brandon Garrett, a Duke University law professor who studies risk assessment tools, said.

David Patton, a chief public defender in New York City, said that “there’s been nothing but confusion” by the BOP with its PATTERN tool. He said prisoners have questions about how they’re being scored “and we have no answers, because BOP doesn’t give us any.”

Using PATTERN to figure out who would qualify as minimum security in light of Attorney General William Barr’s memos to the BOP to step up its efforts to send vulnerable prisoners to home confinement during the pandemic showed the problems with the BOP’s tool. While 20 percent of federal prisoners fall into the minimum security category, only 1.8 percent were identified as qualifying for home confinement under the BOP’s current version of PATTERN.

Blayne Davis is in prison on nonviolent fraud charges at the minimum-security prison camp at the Pollock Federal Correctional Complex in Louisiana, reports. In April 2020, he was scored as minimum risk under PATTERN, and he expected to be released to home confinement as prison staff called names of those who qualified. Davis wasn’t called.

When he pointed to his minimum security score under the chart in the 2019 policy that was publicly disclosed, a staffer handed him a different chart, more restrictive than the one he had. Under the old chart, any prisoner with a score of 21 or less was minimum risk. Davis had a score of just seven. But under the secret chart used by the BOP, any prisoner with a score of six or less is minimum risk. He missed it by one point, he was told, despite his good conduct and having completed 30 classes in prison.

Davis questioned the chart, which he said looked like someone hastily typed up and had some glaring typos. And it wasn’t on official BOP letterhead, like other BOP documents. Staff assured him it was official and that it had been printed off the BOP’s intranet system.

At the urging of Senators Dick Durbin and Chuck Grassley, who co-wrote the First Step Act, the Department of Justice’s Office of the Inspector General (“OIG”) is taking a closer look at the BOP’s efforts to keep more people in prison in the face of Barr’s directives to let them go. They want to know why the BOP has refused to follow the directives, and they’re taking a look at the BOP’s response to the pandemic — which has largely been stall tactics and a failure to respond to serious problems.

But those stall tactics weren’t used to delay releasing celebrity prisoners like former Trump campaign manager Paul Manafort and others, who were promptly freed from prison because of the COVID-19 crisis.