Flight Risk or Misconception? Revisiting and Understanding Absconding Risk

Stephanie R. Penney; Division of Forensic Psychiatry, Centre for Addiction and Mental Health, Campbell Family Mental Health Research Institute, Toronto, Ontario, Canada; Department of Psychiatry, University of Toronto, Toronto, Ontario, Canada

Treena Wilkie; Department of Psychiatry, University of Toronto, Ontario, Canada; Forensic Psychiatry, Complex Care and Recovery Program, Centre for Addiction and Mental Health, Toronto, Ontario, Canada

Alexander I. F. Simpson; Department of Psychiatry, University of Toronto, Toronto, Ontario, Canada; Forensic Psychiatry, Complex Care and Recovery Program, Centre for Addiction and Mental Health, Toronto, Ontario, Canada

At present, there are few validated tools to assist clinicians in assessing absconding risk and formulating viable risk management plans. In this article, we review existing literature on instrument validity and reliability in relation to absconding among patients in forensic care. We examine the predictive validity of a new risk assessment scale for absconding, the Waypoint Elopement Risk Scale-Historical (WERS-H), and assess its incremental utility against a general violence risk assessment instrument (HCR-20V3). Results from all active inpatients in our service (N = 139) revealed 73 individuals who were responsible for 261 absconding events from 2014 to 2020, representing a similar event frequency from a previous census conducted in 2012, but also reflecting considerable annual fluctuations in rate. Confirming results of earlier studies, the presence of substance use and lengthy durations of forensic supervision emerged as key variables associated with absconding. The WERS-H was found to be a significant predictor of future absconding events (incident rate ratio = 1.21, 95% CI [1.07, 1.38], p = .002) and contributed incrementally over the HCR-20V3 Historical scale, suggesting that the use of an absconding-specific risk tool may yield better predictive accuracy.

Absconding; risk assessment; forensic mental health; WERS; HCR-20

“Incidents of escape and absconding from forensic mental health facilities represent significant breaches of public trust and institutional accountability to the parents and families they serve...When considering forensic populations in particular, there is often a heightened perception of risk to public safety, despite accumulating research showing exceptionally low rates of violence and re-offending among these patients while on unauthorized leave...Nevertheless, having reliable data on the base rate of community violence and offending by patients in forensic care while on leave is critical to anchor the oft sensationalized media portrayals of these patients, and to inform policies related to granting leave in a more objective manner...” (p. 80-81).

“The aims of this study were two-fold. First, we present results of an updated census to describe the prevalence of absconding events in a standing forensic inpatient population and to compare the demographic and clinical characteristics of patients who absconded with those who did not...Second, and using the prevalence data from the first aim to establish the base rate of absconding, we sought to validate the WERS-H [Waypoint Elopement Risk Scales; Marshall & Usinger, 2016] by investigating the tool’s predictive and incremental validity vis-a-vis a widely-used risk assessment scheme for general violence. Specifically, we compared the Historical sub-scale of the HCR-20V3 [Historical, clinical, and risk management – 20, version 3] to the WERS-H…We expected to see similar prevalence rates to those reported in recent studies from comparable forensic settings...as well as a decrease in the frequency of absconding from our prior census...Given the limited available data on the WERS-H, our investigation of this tool was exploratory in nature” (p. 82-83).

“Data were drawn from the forensic inpatient population at a large psychiatric hospital in Toronto, Ontario...In regards to the prevalence of absconding, results revealed a similar event frequency from our prior study performed almost 10 years ago, as well as to comparable studies in other Canadian forensic services (Booth et al., 2021; Martin et al., 2018)...As mentioned, there had been several significant changes effected in our program after the first census was completed, including the development of a structured leave application process that was more closely aligned with the risk assessment literature and designed to improve the consistency and transparency of leave decisions...Collectively, these changes reduced absconding in the years following (2014-2016) but with varying sustainability over time...” (p. 83-87).

“...The majority of events identified in this study was brief (<24h) and occurred after the patient had been granted permission to be on hospital grounds or its vicinity. Consistent with other studies from forensic services of low and medium security...patients who absconded from care were not involved in significant violence or re-offending in the community. Patients who absconded were also more likely to have substance use disorders and significantly lengthier durations of supervision under ORB [Ontario Review Board]...the WERS-H allowed for the evaluation of incremental validity against the historical subscale of the HCR-20V3...we found that the WERS-H contributed significantly and incrementally to the prediction of absconding. While the historical subscale of the HCR-20V3 was modestly related to absconding, this contribution did not remain significant when examined alongside the WERS-H. Notably, both the behavioral and clinical items on the WERS-H showed comparable associations with absconding, helping to rebut the notion that only the presence of past absconding predicts future events” (p. 88).

“...Consistent with prior work highlighting the motivating role of boredom and frustration in driving absconding events – feelings which presumably mount over time, it appears that absconsions from forensic services represent a relatively unique phenomenon that may reflect different causal factors over a different timeline. Although expressed feelings of boredom and frustration are difficult issues for teams to address therapeutically, findings highlight the necessity of examining larger systemic and environmental issues so as to create a more transparent and respectful milieu which can contribute to a sense of legitimacy the individual attributes to their detention. On a day-to-day basis, the presence of meaningful and sufficiently challenging daily activities (as opposed to passive ones) has been shown to help ameliorate feelings of boredom in forensic settings...” (p. 88).

“Despite the performance of the WERS-H in this study, an important consideration pertains to the clinical uptake of new assessment tools, and specifically, the potential for burden that is created when introducing multiple tools to appraise risk for different outcomes...If the WERS-H was incorporated into routine clinical practice at our institution, it is anticipated that it could be rapidly completed not only due to its brevity but also given the content overlap with other risk and recovery tools already in practice. Information from the WERS-H could also potentially be used to identify comparatively low risk individuals who may not need ongoing re-assessments of absconding risk at every clinical review period, thereby further increasing efficiencies” (p. 88-89).

“The Waypoint Elopement Risk Scales (WERS; Marshall & Usinger, 2016) are three empirically derived scales (I.e., the WERS Historical [WERS-H], Stable [WERS-S], and Acute [WERS-A] developed to assess risk for absconding among patients (forensic and non-forensic) residing at the Waypoint Center for Mental Health Care in a rural setting...The WERS-H contains eight items detailing the patient’s history of absconding behavior (including attempts to abscond), as well as past problems with substance use and harm to self and others...The WERS-S contains 10 potentially modifiable risk factors for absconding, such as the presence of recent substance use, medication non-adherence, active symptoms of mental illness, and stress...Lastly, the WERS-A is designed to assist in decision-making for off-unit activities, and to highlight any issues that suggest an imminent risk for elopement (i.e., within the next few hours; includes five items documenting recent changes in behavior and personal appearance, presence of anxiety and destabilizers, currently or in the near future)...The developers of the WERS note that any one of the three scales can be used as a standalone indicator of risk for absconding, while acknowledging that a combination of historical and dynamic/proximal risk items can provide more clinically useful guidance of risk and associated management plans” (p. 82).

As always, please join the discussion below if you have thoughts or comments to add!

• Comparing two Primary Methods for Assessing and Managing Violence Risk: Actuarial v SPJ

• Promoting Risk Interventions by Situational Management (PRISM)