Abstract
Recurring carceral evacuations are required for some jurisdictions. Little is known about how these evacuations are planned or executed. This study sought to fill this gap by examining carceral evacuations during disasters in 2020 –2021. Semi-structured interviews were conducted with 24 carceral administrators from three US regions where such evacuations are common, namely the Gulf Coast, the Southwest, and the Northwest. From these interviews, three overarching themes emerged – Carceral administrator prerogative, Interoperability, and Successful end-state – which inform 15 posited recommendations for enhancing carceral evacuation efficacy.
Introduction
Deciding when to evacuate a jail or prison can be difficult, and having limited notice before hand complicates this. For example, the approach of a slow-onset hazard like a hurricane or wildfire provides only a small pre-impact window in which to decide. Disasters like these force carceral administrators to answer several important questions: Should inmates be evacuated before hazard impact? If so, what is the best way to make this happen? If not, are we prepared to shelter in place? On the one hand, researchers have documented the terrible working and living conditions that can result when shelter in place was the decision but preparation was poor (Balaban & Jawetz, 2006; Dement & McAleavy, 2021; Gaillard & Navizet, 2012; Savilonis, 2013; Vumback, 2019). On the other hand, evacuations are costly, logistically challenging, and fraught with safety issues for staff, inmates, and the public (Glick et al., 2013; Kigerl & Hamilton, 2015). The regularity of slow-onset disasters like hurricanes and wildfires in specific regions of the US provided the author an opportunity to study recent inmate evacuations in depth. There is a gap in the literature addressing how carceral facilities execute evacuations. Other than stories which have emerged from media accounts and lawsuits (see Robbins, 2008 for examples from Hurricane Katrina), the actual operations surrounding carceral disaster evacuations are unknown. Indeed, the criminal justice world – carceral facilities included – has been criticized for this very lack of transparency (Chanin & Espinosa, 2015; Geraghty & Velez, 2011). This study aimed to partially address this, thereby advancing the efficacy of carceral evacuations.
Literature Review
By any definition, inmates are readily identifiable as vulnerable persons who are unable to take even the most basic steps for their own safety (Dement & McAleavy, 2021). As a result, they are subordinate to the decisions of prison and jail administrators who are entrusted with their welfare. Without the freedom to act, incarcerated individuals are incapable of coping with emergencies and are arguably more vulnerable than similar non-incarcerated persons (Cannon, 1994; Ferris et al., 2016). Indeed, many inmates are often unaware of the full extent or timeline of hazards as a result of external communication restrictions before and during the event (Dement & McAleavy, 2021). Unlike earthquakes or tsunamis, hurricanes and wildfires in the US occur within generally predictable regions and seasons every year. While no two disasters are alike (Quarantelli & Dynes, 1977), there are enough similarities between them to support planning and preparation. Given this level of predictability, it is reasonable for carceral administrators in these regions to anticipate and prepare for evacuations rather than simply react. Resilience is an important variable in disaster recovery (Aldrich & Meyer, 2014; Boin et al., 2010; Wildavsky, 1988); however, avoiding the disaster altogether by evacuation or mitigation through appropriate preparations beforehand is more desirable (Dement & McAleavy, 2021).
The National Institute of Corrections (NIC) created a disaster planning guide intended for prison administrator use to mitigate the impact of natural and human-induced hazards (Schwartz & Barry, 2009). Shortly before Hurricane Katrina, the NIC, a component of the Federal Bureau of Prisons (BOP), published its guide for preparing for and responding to prison emergencies (Schwartz & Barry, 2009). Savilonis (2013) noted that although it has been in existence for 15 years, awareness of the NIC guide is limited and usage rare. This has resulted in poorly executed carceral evacuations (Dement & McAleavy, 2021). The US Constitution's Eighth Amendment protects citizens from cruel and unusual punishment. The Supreme Court of the United States (SCOTUS) has included maintaining basic humane living conditions (“Estelle v. Gamble,” 1976) and avoiding “deliberate indifference” to inmate safety (“Farmer v. Brennan,” 1994) as duties incumbent on those charged with the care of inmates. At times, post-disaster living conditions in carceral facilities have fallen short of Eighth Amendment mandates (Omorogieva, 2018; Robbins, 2008; Vumback, 2019). Legal scholars have clearly linked the Eighth Amendment to inmate safety both during and after disasters (Omorogieva, 2018; Vumback, 2019). Given this constitutionally mandated duty to care for inmates, the importance of carceral administrators making timely decisions and disaster preparations is clear.
Based on the preceding framework, this study focused on carceral administrators who had been involved in an evacuation during 2020–2021. Although there are scenarios where a low probability, no-notice (FEMA, 2019) event may necessitate an inmate evacuation, this study dealt with the recurring evacuations caused by routine slow-onset hazards. Harm caused by these regular events is not adequately mitigated across the carceral spectrum (Dement & McAleavy, 2021; Omorogieva, 2018; Savilonis, 2013). The discovery of methods which may improve the efficacy of these evacuations therefore constitute a worthwhile undertaking. Accordingly, this study's addressed these research questions:
Research Design and Methods
The proposed research questions suggested a qualitative semi-structured interview and document analysis research design. While awaiting Institutional Review Board (IRB) approval, the author gathered media accounts of disasters that required a carceral evacuation during 2020–2021, building a general base of knowledge about the specific disaster in each jurisdiction. Upon receiving IRB clearance, a small initial sample was purposefully chosen with snowball recruitment following (Mason, 2018). Purposive sampling (Creswell & Creswell, 2018) began by selecting interview candidates with direct administrative control of a carceral facility that experienced a slow-onset disaster and either executed an evacuation, received inmate evacuees, or conducted a-shelter-in-place operation during 2020–2021. Interviewees were assigned an alpha-numeric identifier to ensure confidentiality and all data were stored on a local drive owned and controlled by the author.
With consent from each interviewee, the author conducted 24 1-hour interviews. Of these, 12 involved carceral facilities directly impacted by disaster (hereafter referred to as impacted), six were facilities that hosted evacuated inmates (hereafter referred to as hosts), and six were with state-level Department of Corrections (DOC) executives. Each interview audio was then imported into transcription software. Transcripts were individually reviewed for accuracy then uploaded into qualitative analysis software (QAS) for coding. All transcripts and documents were coded then recoded to ensure coding consistency. This process yielded 62 final codes, 13 categories, and 3 themes, which informed the recommendations in this manuscript. Forty documents were collected for analysis and compared to the interview narratives. All documents were submitted by the interviewees or were publicly available. These consisted of 23 Emergency Operations Plans (EOP)/Incident Action Plans (IAP), 12 Guides, and 5 After Action Reports (AAR). All documents were imported into the same QAS for coding and used to provide triangulation - thus enhancing credibility, reliability, and validity (Creswell & Creswell, 2018; Geraghty & Velez, 2011).
Findings
The findings were drawn from content analysis of the qualitative materials. The author identified 964 quotes tagged with 62 codes after completing the final coding round. The 62 codes were grouped into 13 categories which informed three overarching themes. Each theme is addressed individually.
Theme 1: Carceral Executive Prerogative
One of the earliest themes to emerge was the view of carceral administrators that the status of inmates, whether evacuated or not, is their exclusive prerogative. Although emergency management (EM) authorities were executing evacuation protocols for the public, carceral administrators were universal in the belief that their jurisdiction's Office of Emergency Management (OEM) did not have a role to play in safeguarding inmates.
Planning and Exercises
Few of the carceral facilities interviewed had exercises to prepare for an evacuation. Their frequent response to actual disasters became the de facto exercise.
Staffing Issues
Not only did the disaster tempo in high frequency regions affect the perceived need for regular exercises by carceral administrators, it also impacted their staffing. The warden of a large county jail on the Gulf Coast revealed that keeping adequate staffing levels was difficult because employees were expected to stay and work rather than evacuate with their families. To address this, one warden advanced the jail's readiness by introducing new staff to their internal Emergency Operations Plan (EOP), using a “Hurricane Packet” during employee orientation. The Hurricane Packet is an individual disaster guide of sorts. It informs new staff of what is expected of them during the approach of a hurricane and lists personal protective actions. These measures included having a plan to evacuate one's family without the presence of the staff member, knowing how to turn off water or gas to the home, and developing a supply cache. This was in lieu of disaster exercises because new hires will experience the “real thing” within the next 12 months.
Guide Use
Carceral administrators rarely used a guide like that published by the NIC. This is especially true of facilities that experience frequent disasters and seems counter-intuitive. Administrators reported that because disaster response occurs so frequently in their region, they rely on the experience of their seasoned staff- institutional knowledge rather than a guide.
Theme 2 – Interoperability
Evacuee Tracking
The mass evacuation of inmates has problems similar to those of a public evacuation. One of the most important among these is the need for tracking evacuees. Knowing who is being evacuated and when and to where is essential to managing sustainable services to evacuees and anticipating their return (FEMA, 2019). Using bar-coded bands allows disaster professionals to track not only evacuees, but also their pets, luggage, and durable medical equipment (FEMA, 2019). Tracking inmate evacuees was less developed. Several carceral facilities in this study used a wristband system for inmate identification but there were no barcodes. Instead, bands displayed the inmate's name and photo, and were used internally for corrections officers from different shifts to identify inmates moving between areas within one facility. This created problems when inmates were evacuated as the bands were useless to the host facility. Impacted carceral facilities had to rely on a printed or handwritten manifest to keep track of inmates departing the facility. Carceral administrators who used a handwritten or printed manifest for evacuating inmates spoke about the accountability problems. Not only was creating an accurate manifest difficult when leaving the impacted institution, achieving accountability on the receiving end was just as daunting. One jail commander who evacuated his 1,200 + inmates recounted that 23 different jails across the state were expected to receive them, but the inmate groups were driven to unexpected locations. It took 3 days to locate them. The lack of a standard automated evacuee tracking system further complicated keeping accurate records, even after the initial evacuation was over. Host jails moved impacted inmates between other host jails but did not notify the impacted jail. Moreover, several inmates were improperly released by their host without the knowledge of the impacted jail. This was discovered when the released inmate showed up at the impacted facility seeking personal property. Because county jails are independent of each other, maintaining inmate accountability during movement was difficult. Inmate evacuee tracking was also a problem at the state level. Prisons within a single state used the same computer system, yet still had trouble maintaining accountability of evacuated inmates.
Jail Management Systems
Jail management systems (JMS) in carceral facilities are not interoperable. This is especially true at the county level where each jail is free to choose which JMS to use. Without exception, jail commanders in every sampled state acknowledged that jail management systems are not the same between counties. While there have been attempts by states to standardize county JMS, these fell apart because each county did not keep current with the software. Not having the same version of software prevented neighboring counties, and even facilities in the same jurisdiction, from sharing data from their JMS. Prisons within a single state fared better because they used the same computer system. Although there was no electronic transfer of data during the evacuation, once an accurate count was achieved in the days following impacted inmate arrival at the host facility, inmate movement was documented in the JMS from one prison to the other.
WebEOC
Not having an interoperable JMS was only one technical hurdle that both impacted and host carceral facilities had to overcome. Another was being aware of and leveraging the WebEOC software used by the EM professionals in their jurisdictions. WebEOC is the de facto national web-based emergency operations center software because it is used by FEMA (FEMA, 2018, p. ii). The WebEOC software is created, sold, and maintained by Juvare (Juvare, 2021b). State and local OEMs across the US use WebEOC to achieve situational awareness and handle disaster-related resource requests. Despite its importance in requesting support assets, few carceral administrators have access to WebEOC or are even aware of its existence. One jail commander in a small rural county was fortunate because the OEM consisted of one person who was also a deputy sheriff.
Legal Problems
Having limited or no access to WebEOC is only one of the technical hurdles carceral facilities had to overcome. When inmates were evacuated from the impacted jurisdiction, they were physically separated from legal counsel and the courts. This made adjudication difficult. Host facilities attempted to use virtual platforms like Zoom to allow adjudication to continue. However, rural jails with limited or no broadband access could not support this process. Host institutions had to then arrange transport to a facility where a virtual meeting could take place. Phone conversations were relied upon for simple attorney-inmate consults.
Transportation
Coordinating inmate transportation for a mass evacuation with little notice was difficult. Carceral facilities were accustomed to securely transporting small numbers of inmates during non-emergency periods for court appearances, work details, or housing transfers. However, moving them en masse during an evacuation was more difficult because of security concerns. Keeping inmates securely contained within vehicles required not only more vehicles, but also more persons than could typically be supplied by the impacted facility. Several sampled jurisdictions had agreements with the local school board to use their buses for evacuation or arranged to have impacted inmates picked up by the host facility. Inmate embarkation and debarkation created another set of problems. These events were intended to be staggered to allow time for correctional staff to account for inmates as they exited each bus. Mass departures and arrivals created a bottleneck and contributed to the loss of accountability and maintaining an accurate headcount. Deciding which inmates should be in what evacuation vehicle going to which host institution was another transportation issue. Inmates not allowed to cohabitate were also not permitted to be in the same vehicle. Political posturing for control during the inmate evacuation process was common, especially between counties because there was a peer-to-peer relationship. The county-to-state relationship was less tense. The DOC in one of the sampled states took the lead in arranging evacuation transportation, thus reducing the number of squabbles.
Having rest stops for inmates along the evacuation route was another challenge. The haste with which inmates had to be evacuated before hazard onset revealed the inadequate planning for emergency respite sites. FEMA (2019) defines emergency respite sites as “a location along an evacuation route that can support transportation-assisted evacuees and self-evacuees. Respite sites may include fuel stations, restroom facilities, and access to water” (p.9). Unlike a public evacuation, inmates could not simply get off the bus at a respite site as the risk of escape or other security issues was simply too high. Indeed, this occurred in one prison while evacuating female inmates to a host facility 120 miles away. Wilson (2020) reported the female inmate's account through her attorney, describing the poor conditions on the transport bus when inmates were not allowed to exit the bus for rest stops.
Securing inmates during transport involved more than simply placing officers on the busses and assigning police escort vehicles. Physical restraints limiting movement of the hands and feet were required and in much greater quantities than was routine. Interviewed carceral staff preferred hard restraints like metal handcuffs, ankle binders, and chains instead of plastic locking strips (zip ties). Carceral administrators with a recurring need for evacuations discovered zip ties, although cheaper, to be a greater security risk because of the ease with which they can be broken. When the impacted facility did not have enough hard restraints for their inmate population, the host facility was petitioned for theirs. The construction of hard restraints for use on school buses was another technique several institutions used to overcome shackle shortages. In these instances, individual sets of handcuffs used by road deputies and municipal police were modified and fitted to sections of chain that were run beneath school bus seats. The relatively low-tech nature of restraints and transport vehicles makes this type of equipment interoperability easier to achieve.
Theme 3 – Successful End-State
The outcomes of carceral response to slow-onset disaster were as varied as the impacted jurisdictions. The range of success metrics described by carceral administrators was broad, but there were several consistent replies.
Security
One of the first elements carceral administers described for a successful disaster outcome was security. Administrators consistently responded that both inmate and staff security was necessary. Adding inmate headcount to crowded host facilities raised the likelihood of inmate-on-inmate violence and attacks on staff. Indeed, forcibly adding inmates to a facility beyond its designed capacity required placing them in non-cell holding locations. Much of the unoccupied space at the host facility was used for bed space, including gyms, old records rooms, and unoccupied cells. Keeping inmates with persons with whom they were familiar was a tactic utilized by correctional staff to lower conflict probabilities. Patterns of interaction were already established during time spent in the impacted facility, so keeping that routine at the host facility, as much as possible, reduced the likelihood of misconduct.
Staffing
Staffing availability was another variable carceral administrators in the interviewed regions considered necessary for success. Ensuring staff were present for a carceral evacuation when their own families were evacuating required setting a report-for-duty expectation early. Staff must have a disaster plan for their own families before attending to the needs of inmates. The previously cited Hurricane Packet for new employees was a best practice uncovered by this study. Having a documented plan for family care increased the probability that staff would be able to report to and stay at work. Staffing at host facilities was as much of a concern as that of impacted facilities. Increased carceral staff at host institutions were necessary to properly secure, transport, and monitor the swelling number of inmates. This was easier to achieve for the host than for the impacted facility. One way this was accomplished was to relocate impacted staff to temporary housing close to the host facility if the host was too far from the impact area for a daily commute. Impacted staff simply commuted to closer host facilities, usually a neighboring county jail.
One problem for carceral administrators was how to keep idle staff at impacted facilities employed while there were few, if any, remaining inmates. One emergent solution to alleviate this situation was staff attrition. Carceral staff in the impact area who chose to leave the organization after the evacuation were not replaced until inmates began to return. Another solution was to give staff various jobs to keep them productive. Examples included transporting inmates between facilities, passing out disaster supplies to the public, working as crossing guards, acting as school resource officers, and assuming regular patrol duties.
Wraparound Services
When evacuation was required, carceral administrators indicated that wraparound services at the host facility were key to any kind of successful end state. From FEMA (2021), “Wraparound services include, but are not limited to, sheltering, feeding, medical care, access to distributed commodities, personal assistance services, spiritual care, childcare, clothing assistance, case management, transportation assistance, and vital records assistance” (p.121). Although intended to describe the services available to non-carceral evacuees, most of these still applied to inmates, at least in the minds of the sampled carceral administrators. Having enough food, water, and medicine for both in-house inmates and those they were receiving, was a planning concern for the host facility and one managed in several ways. One method was to have a contract or mutual aid agreement with another facility outlining the supplies impacted inmates would arrive with and ones the host facility would be expected to provide. Adequate bed space at the host facility was another variable considered. Solutions included using so-called “boat beds” that resemble plastic sleds just large enough for bedding material and an inmate. These were laid on the floor along the wall of congregate areas or in cells. Bed space was also found at host institutions if vacancies existed in active facilities or idle buildings. Shuffling their resident inmates to other areas opened beds for evacuees while keeping the two populations separated. Although the proximity of inmates to each other during the COVID pandemic violated the CDC's recommendations for safe distancing, the alternative was not a viable option.
Regarding how COVID influenced the evacuation decision, administrators uniformly saw approaching fire, hurricane, or chemical spills as a greater risk than the virus. Indeed, some protective actions required greater exposure to the virus rather than less. This was also true during evacuation transport. The time required to evacuate inmates would have been too great to avoid the hazard if social distancing constraints were adhered to on vehicles. Carceral administrators also listed access to legal counsel as one of the wraparound services required to consider evacuation a success. To that end, efforts were made by both host and impacted facilities to ensure some type of inmate virtual access. In situations where this was not available, proceedings were rescheduled or inmates were shuttled back to the impacted jurisdiction for an in-person appearance.
Funding
Evacuating and caring for inmates is expensive. Transportation, security measures, and wraparound services are not free. Sampled jurisdictions financed their operations in several ways. First, jail funding often began outside the carceral facility with EMs who had access to FEMA funds. For example, FEMA dollars were used to purchase additional equipment and pay for staff overtime. Although not a direct revenue stream for local jails, FEMA dollars lessened the burden on the county budget of paying an inmate per diem to host facilities. “Costs related to providing emergency assistance such as mass care may be reimbursable through FEMA's Public Assistance (PA) program. Reimbursement would be applicable to cover costs that support emergency protective measures” (FEMA, 2009, p. 62). Second, sheriffs’ offices faced with routine evacuations drew on experience and public support to gather funds independent of FEMA. One jail commander praised the sheriff for getting a special emergency fund passed by county voters that was used to cover disaster-related expenses. Voter support was made possible by the repeated evacuations the county had experienced. While FEMA funds did assist county jails with recovery, this process was lengthy and prevented FEMA dollars from being applied to immediate needs. Planning to be independent of FEMA funds in the short term was necessary. Third, many counties in the interviewed regions had more than one jail that housed federal or state inmates. Housing state and federal inmates generated jail revenue because the bed-per-day fee from the state DOC and the federal BOP was greater than the daily charge for county inmates. For example, one jail reported receiving $3 per day for a county inmate and $26.50 per day for a state inmate. This higher state per-day rate did have a downside, as it was the rate the state's host institutions expected to be paid by the impacted jurisdiction during inmate evacuations. Understandably, host jails across the state were more than willing to take these evacuated inmates. Unfortunately, the loss of state or federal inmate revenues to impacted jails during an evacuation depleted general county funds, denying the Sheriff a revenue stream as counties had to pay a host facility the full state rate. Additionally, extensive damages to the impacted facility necessitated a lengthy and expensive use of host facilities.
Finally, jails utilized work programs as a revenue source independent of state or federal oversight. One impacted county jail found regulations surrounding housing state or federal inmates too cumbersome. They opted to cultivate revenue and reduce costs with inmate programming building furniture to sell and raising produce on a nearby farm. This impacted jail made the decision not to house either state or federal inmates for revenue because of the additional rules and oversight that accompanied housing them. Financial independence was easier to achieve for prisons than for local jails. In contrast to stand-alone county jails, state prisons did not need to rely on external facilities to hold evacuated inmates. Bed space and resources were shared across internal institutions. As a result, when inmates were evacuated from one state facility to another, dollars were transferred between ledgers within the organization rather than paid to a third party.
Discussion and Recommendations
Discussion and recommendations are structured around the research questions. These recommendations are intended to improve the efficacy of carceral evacuations.
RQ1: What are administrators doing to prepare carceral facilities for disasters?
Administrators should be involving OEM evacuation experts in planning. Carceral administrators view their evacuations as a stand-alone venture independent of the OEM. As a result, the assistance to be gained from the evacuation expertise and experience of OEMs is lost. Evacuations are not a core function of carceral facilities, but they are for OEMs. Carceral facilities need to become involved in regular disaster exercises and training events held by the local OEM or create their own and invite the OEM as a consultant. Administrators need to strengthen their understanding of WebEOC. WebEOC usage is vital to improving carceral disaster efficacy. However, the findings revealed that carceral professionals have limited, if any, knowledge of WebEOC or its capabilities. This lack of understanding restricted their ability to fully access available resources. By working with local OEMs, carceral administrators can become WebEOC “smart.” Administrators need to make evacuation decisions early. The time necessary to evacuate carceral facilities is greater than that of the general population because of secure transport demands. Facilities who make an evacuation decision early have more time to prepare and therefore a greater chance of maintaining security and accountability. If an early evacuation was ultimately not necessary, it was at least a rehearsal for those times when it will be. Administrators should cultivate multiple funding sources. Disasters are expensive and the most resilient agencies do not rely solely on funding from typical taxes to cover the bill. Having additional funds helps pay for staff overtime, host facility per diem rates, equipment purchases, and wraparound services. One method of achieving this is to have more than one jail, or many vacant beds. Additional bed space gives jails an opportunity to house state and federal inmates for a premium fee. However, housing these inmates comes with additional oversight and rules that may be undesirable. Additional funding can also be sourced through special taxes earmarked for the disaster needs of the sheriff's office. Inmate labor is another funding stream that can be found in workshops and farms of all types. Products built or grown can be sold to augment revenue or simply used to reduce cost. Administrators must improve inmate evacuee tracking. Findings indicated that a successful evacuation must include end-to-end inmate accountability. Not knowing which inmates are where for days creates problems. For example, families do not know how to reach their incarcerated loved ones, the legal process is interrupted, and losing accountability of high-risk inmates is a danger to public safety. Jails and prisons could adopt Juvare's EMTrack software, or another like it, to maintain accountability. EMTrack is a software package closely associated with WebEOC – also developed by Juvare – that can be used by carceral facilities to overcome JMS interoperability problems. Although developed primarily for hospital patient tracking (Juvare, 2021a), EMTrack can be easily adapted for carceral use. Indeed, the author met virtually with Juvare's EMtrack team several times and observed its capabilities and potential for carceral adaptation. Any type of bar-coded wrist band can be used by the software. This is important because OEMs who regularly evacuate are already using bar-coded wristbands of various types. As a result, finding a ready supply for carceral use would not be difficult. Administrators should resolve JMS interoperability problems. Sampled administrators related how inmate headcount, location, and adjudication status cannot be shared between facilities. Indeed, accurate headcounts and inmate location can take days to achieve after evacuation. Additionally, inmates have been mistakenly released and had to be recaptured. A common JMS system needs to be adopted or a bridge between systems identified. EMTrack, or similar software, has the potential to resolve JMS problems by bridging incompatible systems with a temporary means of inmate tracking and status communication. Like WebEOC is to the nation's OEMs, if EMTrack type software can become the supported standard for carceral facilities within a state or region, accountability is more easily achieved. 7. Using Employee Hurricane Packets is a model for others. Employee Hurricane Packets – although applied and intended for jail staff on the Gulf Coast – can be adopted by other regions. Wildfires, severe winter weather, and flooding are instances where an employee disaster prep packet is useful. Not only are the families of carceral staff better equipped to face the hazard after completing preparations, but expectations of staff duties are also better communicated, staff are free to focus on the job at hand, and the probability of their being at work is improved. 8. Use facility maintenance as a chance to rehearse inmate evacuations. Small jails occasionally need to evacuate for maintenance reasons because there are no available vacant cells for inmates. This is a rehearsal for evacuation operations and provides an opportunity to renew agreements with host facilities. Larger jails or prisons can adopt a similar practice. For example, rather than moving inmates from one pod to another for a plumbing issue, use the opportunity to conduct a small-scale evacuation with fewer time constraints. Lessons learned from this exercise could be applied to the inevitable full evacuation. Admittedly, the funds and coordination necessary for a maintenance evacuation are an obstacle. However, if the maintenance is routine, it can be planned for, and the cost built into the training or maintenance budget. 9. Administrators can use and expand virtual platforms for inmate legal services. When impacted inmates need to be evacuated for extended periods, their adjudication and legal services must continue. Host and impacted facilities are more successful at maintaining continuity of operations when it is possible to conduct this business using virtual platforms rather than expending resources to shuttle inmates. 10. Administrators can use impacted carceral staff elsewhere. Impacted staff left behind after an evacuation can be employed elsewhere. At the county level, staff remain sworn deputies so can assume other duties easily. Prison staff may have similar options. If inmates remain with a host facility for an extended time, not hiring carceral staff lost to attrition can reduce underutilized headcount. When appropriate, a controlled return of inmate evacuees will allow time for a gradual replacement of lost staff. 11. Administrators should evacuate inmates by housing unit. Maintaining the integrity of inmate housing groups during an evacuation promotes routine behavior because groups are used to living together. Prohibitions against certain persons being moved together are easier to adhere to when housing units are already set up for this. Established relationships with cohabitants facilitate familiar behaviors thus making the evacuation process more predictable. Moreover, staff of impacted facilities are accustomed to seeing certain groups of persons together so an unfamiliar face can be an indicator of an accountability or security problem. 12. Administrators should coordinate transportation through the state DOC. The DOC has access to more transportation assets than any single county. These include prison buses equipped for the security needs of inmate transport. Carceral evacuations can be instituted more quickly when these assets are made available before impact. The DOC can pre-stage transport assets if an evacuation appears likely for one or more jails. This is akin to what happens for evacuations of the public. With the DOC coordinating inmate transportation, inter-county political conflicts and confusion can be reduced. The DOC becomes the clear authority and decision maker. If jail transportation assets are used instead of prison buses, the DOC can still provide other equipment including shackles, drivers, and escorts. Moreover, the DOC has the potential to encourage compliance with established plans. 13. Administrators can reduce headcount using techniques developed for COVID. The need to slow the spread of COVID drove efforts to lower inmate headcount through early release. This necessity was the genesis for providing headcount reduction decision criteria. Persons selected for early release included non-violent offenders or offenders close to the end of their sentence. These same criteria could be used to reduce headcounts before an evacuation, assuming the courts approve. A reduced headcount simplifies carceral evacuations by lowering the transportation needs, and number of staff required, exposing fewer inmates to risk, and saving the costs of wraparound services. 14. Administrators should create and use a functioning disaster guide. Sampled carceral administrators are using disaster guides infrequently or not at all. Instead, they rely on the experience and memory of their staff to conduct an effective evacuation. This method is unreliable. When attrition or an election causes the turnover of key leadership, institutional knowledge is lost, and the facility is left to rediscover the same lessons. One of the sampled jails experienced this firsthand. Newly elected sheriff's jail commanders and chief deputies had all been on the job less than a year and had not yet been through an evacuation in their new roles. As a result, when disaster struck, operational relationships did not exist inside or outside the agency and evacuation efficacy was reduced. 15. Administrators should find opportunities to share guides broadly. Several carceral administrators mentioned national and regional corrections associations that publish guides and reports. Agencies like the National Sheriff's Association or National Institute of Corrections are excellent candidates for promoting the collection and dissemination of best practices. However, these guides were primarily accessed by carceral administrators for prison emergencies that did not require an evacuation. None of the sampled administrators shared evacuation-related best practices with these organizations to be passed on to others. Indeed, jail commanders lamented this when the author provided descriptions of the best practices from their peers across the sampled regions. These practices were seen as valuable because of their capacity to reduce problems, but they are not being shared through any medium or enshrined in a useable guide. National corrections and law enforcement associations like the NSA exist for this purpose, but they do not appear to be as effective as they could be. Because many of the recurring slow-onset hazards are regional, individual states may be in a better position to create policy encouraging organizational sharing by illuminating their shared risk.
RQ2: What disaster response methods have yielded successful outcomes?
RQ3: To what extent are carceral administrators aware of and using guides to prepare for and execute evacuations?
Limitations and Future Research
Sample Size
Given the sample size, it is acknowledged that the findings are not representative or generalizable to the entire country (Maxwell, 2002). However, the population size of carceral facilities exposed to routine, slow-onset hazards is much smaller – thus the findings are more applicable to this group. Considering the size of the target population for this qualitative study, the sample size was carefully weighed to achieve the study's aims. Indeed, Patton (2002) argued that “sample size depends on what you want to know, the purpose of the inquiry, what's at stake, what will be useful, what will have credibility, and what can be done with available time and resources” (p. 244).
Geography and Jurisdiction
This study excluded US federal prisons and sampled only state and county carceral facilities that conduct regular evacuations. Facilities with a reduced need for recurring evacuations will find less value in the recommendations. However, several of them can be applied broadly regardless of circumstances. Juvenile detention facilities were also not included in the purposive sample, although they must prepare for and respond to disasters as well. As a result, this investigation's findings and recommendations cannot be generalized to either the federal or juvenile carceral systems. An international comparison of carceral disaster operations is a logical future step. The rise in frequency and scale of hazards demands more anticipation and better preparation around the globe (Hore et al., 2018). Those nations with more centralized governments than the US may provide lessons on the advantages of standardization and interoperability. What is learned about the preparedness of carceral systems internationally can be shared to improve the preparedness of all.
Inmates as Interviewees
This study explored how carceral administrators prepare for and respond to the impact of slow-onset hazards. Understanding the experiences of the impacted inmates would add a valuable – and relatively unexplored – perspective. The press has published a few inmate accounts of their disaster experience but there is currently no empirical research. Addressing this gap in the literature would require careful ethical consideration and adherence to the requirements of the IRB. Best practices concerning the use of inmate labor in disasters need further study as well; most state planning documents include some reference of inmate labor during times of disaster (Purdum & Meyer, 2020). Although not the focus of this study, the findings highlighted several methods for the effective use of inmate labor during disasters at both the county and state levels. Which of these is most cost-effective, the supervision needed, and how inmates are selected for participation are all areas in need of further research.
Conclusion
This study sought to critically evaluate carceral facilities preparedness for and response to routine, slow-onset hazards. Thousands of impacted inmates were evacuated or sheltered in place during 2020–2021. However, little is known about what planning preceded this, how it was accomplished, and what problems were encountered, demonstrating the need for this study and research focus. This includes determining (1) what are administrators doing to prepare carceral facilities for disasters; (2) what disaster response methods have yielded successful outcomes; (3) to what extent are carceral administrators aware of and using guides to prepare for and execute evacuations. Interviews with 24 carceral administrators who lived through disasters in 2020–2021 and routinely evacuate their facilities in the US Gulf Coast, Southwest, and Northwest were conducted and analyzed, along with supporting documents. Three themes emerged which informed 15 recommendations for improving carceral evacuation efficacy. Several of these recommendations are simply best practices that can be adopted by facilities with little effort. Others will require changes to policy and the collective effort of the carceral body within states.
Key Findings
The three most urgent recommendations are (1) involve the OEM – the evacuation experts – in planning; (2) resolve interoperability issues to improve inmate evacuee tracking and data transfer between facilities; (3) create, use, and share disaster guides. First, for all disaster phases, carceral administrators and EMs did not envision a role for EM in carceral disaster operations. Although the disaster preparation and response of these two occurred at the same time (parallel), they did not overlap (synchronous) (Berlin & Carlström, 2011). Participating carceral organizations viewed their OEM counterparts as unhelpful with the practical needs of their disaster operations and unconcerned with the goings on in the jail or prison. When carceral organizations decided to evacuate or shelter in place, they did so independent of the OEM. There was no consultation with the OEM regarding what to do or when to do it. Second, interoperability between carceral entities was limited because of incompatible jail management software (JMS). Information sharing between impacted and host facilities was reduced to an ineffective paper-based and email approach. This included inmate legal status, demographic information, medical condition, and housing assignment. Additionally, lack of a common JMS made inmate tracking during an evacuation problematic. Interoperability was also constrained between carceral organizations and OEMs. However, a knowledge deficiency rather than a technical one is to blame. Carceral administrators were unaware of, or unfamiliar with, the capabilities of WebEOC and its role in general disaster operations. This shortfall prevented wardens and jail commanders from tapping the resources available to them through the OEM. Improving this knowledge can help make heretofore unknown resources available to carceral executives and promote a relationship with the OEM; WebEOC account holders are typically OEM employees.
Finally, rather than using a premade disaster guide or creating their own, carceral administrators in this study's high-risk regions relied on staff institutional knowledge developed through repeated exposure to the same hazards. This was problematic because when staff left the organization, their knowledge went with them. Because little documentation recording disaster actions existed, there was limited sharing of disaster best practices between carceral entities. The National Institute of Corrections and the National Sheriffs’ Association are intended as a clearing house for carceral information, but this does not occur for disaster operations. This situation was lamented by carceral administrators who felt they could have benefited greatly from the disaster experiences of their counterparts. However, these same administrators struggled to document or share their own best practices. The SCOTUS has secured Eighth Amendment protections for inmates during disaster. In the past, conditions endured by inmates have been inhumane, unnecessary, and preventable. Inmate suffering during disasters continues although a path to reduce it exists. The findings of this study demonstrate a clear need for carceral entities to learn from past disaster response mistakes. Adopting the recommendations posited here will engender greater preparedness and help ensure the well-being of an inherently vulnerable population for whom we all have a duty of care.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article