Background

A few years ago, the Pennsylvania Department of Corrections experienced a cluster of suicides within a short period of time. After each suicide, PDOC adhered to our standard suicide clinical review process in an effort to identify areas of improvement or needed remediation. Our psychology office also reviewed the cluster of suicides together as a whole to identify any broader systemic concerns that may have occurred.

In this cluster review, we identified that the percentage of individuals categorized as “double celled” at the time of their death—meaning they had a cellmate assigned to their cell—appeared high based on our previous experience reviewing and understanding suicides. Consequently, we re-reviewed each suicide within the cluster and discovered that in fact only one of them was technically double celled at the time of the suicide; that is, in only one instance was the cellmate present in the cell when the decedent was discovered.

In the other four cases, although the individuals were categorized as double celled, they did not initiate the suicide until their cellmate was away or had exited the cell. The individuals were actually alone in the cell by themselves at the time they initiated their suicide.

Upon discovering this inadvertent data collection error, we initiated a larger retrospective review of all suicides that had occurred within PDOC since 2000 in an effort to clarify the precise housing status of each decedent at the time of their discovery. Looked at through the lens of our new understanding of the concept of being double celled versus being alone, our review of this larger dataset revealed the same error in our understanding and categorization of housing status. The result was staggering: in 95% of all suicides that have occurred within the PDOC since 2000—174 of 184—the individual was alone in a cell at the time of the suicide.

experience higher rates of mental illness and serious mental illnesses than men. SCI-Laurel Highlands specializes in delivering the highest level of acute medical care in our system, including care for people who are terminally ill or near end of life. SCI-Waymart is responsible for delivering PDOC’s highest level of inpatient mental health care and specializes in housing our most seriously mentally ill male individuals. Despite high-suicide-risk patient populations, those four SCIs, looked at together, had only ever experienced two suicides.

That finding was counterintuitive to what we thought we knew about suicide risk. How were those institutions, which house apparently higher-risk populations, having so much success at preventing suicides?

We informally interviewed staff from each of the prisons and asked, “What are you doing differently?” Their answers were consistent: “We’ve learned how to work effectively with these populations. We know how to keep them safe. We treat them professionally and humanely; we speak to them and treat them with respect.”

That seemed like a plausible explanation, but it didn’t quite fit with what the data were telling us. While we agreed that our staff at these institutions were professional, we thought there may be something more going on, and in fact there was.

At each of these four prisons, there are very few cells. Most of their physical plants are essentially open-dorm style settings. Most individuals are housed in large open areas, visible to many other people, which creates infrequent “alone time.” In addition to their excellent staff, one potential reason these prisons had so much success in preventing suicides was that the individuals in these settings were rarely housed alone.

Our next step was to critically review our operational policies and practices. We began with a review of our Z-code policy, which outlined operational standards and guidelines for single and double celling. We discovered that our Z-code policy indicated that having mental health problems or a history of being dangerous toward self, self-mutilative, or unable to care for self were acceptable singular reasons to consider housing someone in a single cell. Our data, however, suggested that those reasons, taken alone, were likely contraindicated for being housed alone.

As a result, we took immediate action and issued a memo to the organization revising the Z-code policy to prohibit assigning Z-codes for those contraindicated reasons. Additionally, we directed that all SCIs commence meaningful reviews of all individuals single celled at that time to determine whether the individual could be safely double celled.

Other improvements we have implemented, based on this suicide data review:

• Increased the frequency of security rounds on all Restrictive Housing Units and Special Management Housing Units statewide, from once every 30 minutes to unpredictable intervals with no more than 15 minutes between checks, with special emphasis on those individuals housed alone. By increasing the frequency of security rounds, we decrease the amount of time that people who are housed alone are alone.