Abstract
Granted leave is an integral part of forensic-psychiatric treatment. To prevent serious crimes or absconding during leave, careful assessment of risk of such misconduct is necessary. This check is carried out by a decision-making process, which considers patients’ freedom, clinical benefits, and the need for public safety. A systematic examination of processes surrounding granted leave can help decision-makers to proceed in more organized and transparent manner. Representatives of all professional groups at 12 forensic-psychiatric hospitals in Bavaria were questioned using semi-structured interviews. Information on granted leave was quantified using a self-constructed checklist and descriptively evaluated. The extent to which legal requirements and other process characteristics are implemented was examined. The implementation of requirements varied across hospitals in terms of organizational and personal framework conditions. A joint agreement among all hospitals regarding the definition and handling of misconduct during granted leave, revocation of granted leave, and risk assessment addressing absconding would be particularly desirable. Common standards of excellence could contribute to more transparency and strengthen the trust of patients and the public in the approach of a hospital.
In forensic-psychiatric treatment granted leave is an important part of the patient rehabilitation process. In Germany, granted leave includes being allowed to leave the hospital ward or hospital grounds first escorted by staff, then alone and up to community access with overnight stay for several days. This gradual release of restrictions on freedom until discharge improves patients’ quality of life, increases their therapy motivation, strengthens therapeutic alliance, and enables testing of learned coping strategies under real-world conditions (Hilterman et al., 2011; Lyall & Bartlett, 2010; Schel et al., 2015; Walker et al., 2013). However, granted leave also harbors a risk of absconding (with or without reoffending). These incidents are rare but could cause serious harm to others (Scott & Meehan, 2017; Simpson et al., 2015; Watson & Choo, 2021). Therefore, granted leave decisions must address patients’ needs and public safety at the same time. Helpful in this context would be standards of excellence, which can improve transparent and coherent decision-making about granted leave, but also reassure those who must make decisions about granted leave while facing the dilemma between patient rights and community rights (Birgden, 2008; Völlm et al., 2018). Kennedy (2022) proposed such standards of excellence for forensic-psychiatric hospitals. According to these, a hospital’s model of care should be consistent with legal requirements and, among other things, contain process descriptions that address granted leave decisions and a plan of action in emergencies such as absconding. For example, care models should include periodic reviews of patients’ therapeutic progress and risk prior to decision-making about granted leave. Decision-making should follow a multi-level process involving experts involved in treatment. However, it has been shown that legal requirements for forensic-psychiatric treatment vary across and even within countries and are implemented differently from hospital to hospital (Barlow & Dickens, 2018; Edworthy et al., 2016; Hilterman et al., 2011; Neumann et al., 2019).
German regulations and procedures surrounding granted leave
According to German Criminal Code (Strafgesetzbuch; StGB, 2022) individuals can be sentenced to a mandatory treatment order in a forensic-psychiatric hospital if (1) they have committed any offense for which they are found diminished responsible or fully irresponsible owing to severe mental disorder, and (2) are at high risk of reoffence. In most cases, treatment is provided in two therapeutic sub-areas (addictive disorders under Section 64 of the Criminal Code; other mental disorders under Section 63 of the Criminal Code), and in secure inpatient forensic-psychiatric hospitals consisting of wards with various levels of security (low, medium, high, etc.).
According to the legal requirements for forensic-psychiatric treatment in Bavaria (Bayerisches Maßregelvollzuggesetz; BayMRVG, 2020) and related administrative regulations (Verwaltungsvorschriften, VVBayMRVG), granted leave transfers a patient from higher to lower security levels within one hospital. In this process, each hospital is run entirely independently. Once a leave is granted, it remains unchanged until further notice. The main levels of granted leave are: Level A—escorted leave of high security area or hospital grounds; Level B—unescorted leave of high security area; Level C - unescorted leave of hospital grounds; Level D - community access with overnight stay. Granted leave can be combined with supervision instructions such as temporal or spatial constraints. It can also be suspended or revoked if new circumstances such as absconding, or an increase of symptoms appear. Decisions about granted leave should be the result of a multilevel process carried out by clinicians directly involved in the treatment. At least at one of those levels (e. g., on ward, across wards), a multi-disciplinary decision-body should take place. An additional clinician not involved in the treatment must participate at decisions about patients who committed crimes such as murder, arson, or sexual offenses. Ultimately, however, the chief psychiatrist has the exclusive right of decision whether to approve a granted leave. There are no other approval boards involved in decisions about granted leave in Germany. Decisions about high-risk patients, who have committed serious crimes (e.g., murder, arson, sexual offenses), include an assessment of clinicians not involved in the treatment.
Treatment planning must be updated at least every six months, and, in line with this, also the indication for granted leave checked. Before granting a leave, clinicians must assess the therapeutic progress and the risk of absconding or reoffending. The legal requirements also recommend that certain risk factors and their changes be considered in risk assessment. Those are delinquency, predispositions, mental condition, inpatient behavior, and social environment. Risk of reoffending and absconding must be assessed separately. The use of risk assessment tools or the internal hospital handling of absconding and reoffending are not regulated by BayMRVG (2020).
Studies on granted leave
To date little is known about how patients progress from one level of granted leave to the next. Several studies addressed the use of actuarial and structured professional risk assessment approaches in the context of decisions about granted leave (de Vries Robbé et al., 2011; Hilterman et al., 2011; Müller-Isberner et al., 2007). Others focused on the patient characteristics for decisions (Doyle et al., 2012; Stübner et al., 2006). There are only few studies, that describe the process up to a decision about granted leave, and their results are not generalizable given the varying national and international legal conditions (Barlow & Dickens, 2018; Neumann et al., 2019; Stübner et al., 2006).
A review by Barlow and Dickens (2018) compiles various process characteristics and suggests inconsistencies in care models: patients, hospital authorities or clinicians can initiate a decision about granted leave. An application is usually followed by multidisciplinary clinical team discussions. In many institutions, the experience of nursing staff appears to be of great importance in preparing a decision to grant leave, but there is no standardized procedure for considering nursing staffs’ input. Leave can ultimately be granted by the clinical team, but also by the chief psychiatrist, administrations, or external approval boards.
Stübner et al. (2006) found, that less than half of all Bavarian (Germany) forensic-psychiatric hospitals used a valid risk assessment tool for leave decision-making. Instead, clinicians relied on unstructured clinical judgment or used self-constructed checklists listing risk factors selected from various established assessment tools. Consequently, Stübner et al. (2006) found considerable variation between hospitals in terms of which risk factors they considered. However, no further process description was given in this study.
Neumann et al. (2019) examined the process of granting leave in forensic-psychiatric hospitals of another German federal state and found inconsistencies: In some hospitals, the initiative for requesting a leave came exclusively from the patients; in others, consultation with the responsible therapist or with the entire treatment team of a ward was required first. Also, the number of decision-making-levels, professionals involved, and the extent to how patients were interviewed during this process varied. Risk factors considered as important for decisions about granted leave differed across hospitals, and mandatory standards regarding valid risk assessment tools were absent (Neumann et al., 2019).
Purpose of the present study
This study describes the processes of decision-making about granted leave in Bavarian (Germany) forensic-psychiatric hospitals and its underlying legal requirements. Since only one study has specifically addressed these processes in German forensic-psychiatric hospitals (Neumann et al., 2019), the present study aimed to expand the national database by an evaluation of several process characteristics in another federal state. A methodological approach of an interview was chosen to be as unbiased and open-ended as possible. It has been shown that expert practitioners can provide accurate information on their own work steps (Stübner et al., 2006). The extent to which the requirements of state law (BayMRVG) for granting leave were implemented was examined, and all hospitals were compared descriptively. We additionally examined few process characteristics not included in legal requirements but described in previous studies (Barlow & Dickens, 2018; Lyall & Bartlett, 2010; Neumann et al., 2019; Stübner et al., 2006).
Method
Measures
Data on processes surrounding granted leave were collected through interviews involving eight open key questions: Please describe the levels of granted leave at your hospital; Please describe how decisions about granted leave are made at your hospital; Does the process differ for certain patient groups?; What are the reasons for downgrading, suspending, or revoking granted leave?; Is there a documented in-hospital agreement for granting leave and/or revocation of granted leave?; How do decision-making-bodies look like?; How is absconding defined?; Please describe how you handle absconding. A total of 24 pre-formulated supplementary questions were used to increase the inter-individual survey objectivity (examples: Who initiates decision-making about granted leave?; Which professional groups are involved in the decision-making process?; Who takes the final decision?). Since all interview answers were given in a narrative style, they were quantified using a checklist developed for this purpose. The coding-categories of the checklist were designed to determine the implementation of the legal requirements and to evaluate further process characteristics (see Sklenarova et al., 2020). Implementation of legal requirements was assessed via 22 items: nine items allowed a dichotomous rating (0 = legal requirement not implemented; 1 = legal requirement implemented); seven items a Likert-scaled frequency rating (1 = legal requirement not implemented; 2 = legal requirement rarely implemented; 3 = legal requirement often implemented; 4 = legal requirement always implemented); and six items had individual four-level Likert scales (e.g., 1 = no documented in-hospital agreement on granted leave; 2 = documented in-hospital agreement on granted leave is imprecise; 3 = in-hospital agreement is documented but not transparent for patients; 4 = in-hospital agreement is documented and transparent for patients). For this study, a cutoff was set at 2, which indicated implementation of legal requirements and additional process characteristics (for list of items, see Tables 1 and 2). There were two kinds of missing values: it was not possible to gather the information from the interview (“missing information”), or the participant has explicitly declared not being able to answer the interview question (“don’t know”).
Proportion of affirmed implementation of legal requirements and process characteristics of granted leave decisions.
Note. N = 133. Proportion of affirmation (“legal requirement / characteristic is implemented”), negation (“legal requirement / characteristic is not implemented”), missing knowledge (“Don’t know”), missing data (“Missing”).
Adverse incidents = insult, attempted or completed suicide, assault on or by patient, possession of prohibited items, substance abuse.
Patient who has committed serious crimes (murder, arson etc).
Therapeutic sub-areas = addictive disorders under Section 64 of the Criminal Code vs. mental disorders under Section 63 of the Criminal Code.
Legal requirement / characteristic is often or always implemented.
Implementation of legal requirements and process characteristics of granted leave decisions affirmed by the majority of professionals per hospital.
Note. N = 133.
Adverse incidents = insult, attempted or completed suicide, assault on or by patient, possession of prohibited items, substance abuse.
Patient who has committed serious crimes (murder, arson etc).
Therapeutic sub-areas = addictive disorders under Section 64 of the Criminal Code vs. mental disorders under Section 63 of the Criminal Code.
Legal requirement / characteristic is often or always implemented.
At first leave-level only.
Procedure
The descriptive cross-sectional study was part of the project "Decisions on the gradual granted leave in the Bavarian forensic-psychiatric hospitals", which has already been described in more detail by Sklenarova et al. (2020). The project was approved by the ethics committee of the DGPs (German Society for Psychology e.V.) (2019-10-18VA).
Participants were recruited from 12 Bavarian forensic-psychiatric hospitals for semi-structured interviews using a nonprobabilistic conscious selection process (Etikan et al., 2016). Selection criteria were at least 1 year of professional experience at the forensic-psychiatric hospital and participation at decisions about granted leave. All participants provided written informed consent to participate in this study. Data were collected between June 2019 and March 2020 using a combination of pseudo- and anonymization. Therefore, while the participation was anonymous, it was possible to assign the participants to a particular hospital via pseudonyms.
A total of 133 professionals from 12 hospitals participated in this survey all involved in decision-making. More than half of the participants were male (n = 74, 55.6%). At the time of the interview, the participants had worked in the hospital for an average of 12 years (SD = 8.75, 1-33 years). The sample consisted of 12 chief psychiatrists (9%), 11 security officers (8.3%), and representatives of the professional groups involved in the treatment: 23 (17.3%) psychologists, 23 (17.3%) nurses, 23 (17.3%) social workers, 21 (15.8%) psychiatrists, and 20 (15%) professionals from various free complementary therapies (occupational, sports, music therapy, etc.).
Interviewers took notes during the interview and transferred them retrospectively to a digitized interview protocol. Subsequently, a member of the research team other than the interviewer quantified the data using the checklist (Fakis et al., 2014). The interrater reliability of the measure was excellent (ICC (2,3) = .974, 95% CI[.964; .981]). Finally, all hospitals were compared in terms of the extent to which they implemented legal requirements and other process characteristics. A requirement/characteristic was classified as implemented if it was affirmed by >50% of the professionals at each hospital. To support the comparative analysis, additional interview statements were selected and presented as data. It was assumed that the interview excerpts would describe each hospital’s decision-making process in more detail. A characteristic was added as descriptive of a process if it was mentioned at least once in each hospital (see process characteristics in Table 3).
Implementation of additional process characteristics of granted leave decisions affirmed by at least one professional per hospital.
Note. N = 133.
At first leave-level only.
For high risk patients who have committed particular serious crimes (murder, arson etc).
For patients with addictive disorders under Section 64 of the German Criminal Code only.
For patients with mental disorders under Section 63 of the Criminal German Code only.
Statistics
Descriptive analyses (absolute and relative frequencies) were performed. All statistical analyses were performed using the IBM SPSS Statistics (Version 25), IBM Corp (2017).
Results
Data on the implementation of legal requirements and other process characteristics are presented using the following defined decision-making process steps: preparation of decision about granted leave, decision-making-body, decision about granted leave, and revocation of granted leave. Table 1 shows the proportion of professionals affirming implementation and the proportion of both types of missing values. To simplify a comparative description of internal hospital processes, no tabular representation of absolute and relative frequencies is provided. Table 2 instead graphically illustrates implementation by hospitals using check marks. Of the legal requirements only three hospitals implemented more than 80% of the process characteristics. The nine additional, from legal requirements independent process characteristics are presented in Table 3.
Discussion
This study examined the processes surrounding decisions about granted leave in 12 Bavarian forensic-psychiatric hospitals based on legal requirements. Overall, interview statements from study participants revealed that most of the legal requirements are implemented. However, the hospitals differ in how they proceed in decision-making about granted leave in detail. Similar results have already been described in previous studies (Barlow & Dickens, 2018; Neumann et al., 2019; Stübner et al., 2006).
During the preparation phase for a decision, a patient applies for granted leave. Some hospitals combine this application with special tasks for the patients (e.g., self-reflection, or in-person application presentation). Both stimulates patients’ active involvement into the treatment and provides preliminary evidence to suggest that the hospitals are applying shared decision-making policies during decisions about granted leave. Such policies were found to support patient autonomy, improve the therapeutic relationship, and thus lead to better treatment outcomes (Swift et al., 2018; Trusty et al., 2019, Tryon et al., 2018). In addition, about half of hospitals survey patients as part of the decision-making-body. This does not necessarily mean, that the decisions at the other half are made without active patients’ participation as they always initiate decision-making about granted leave. Also, the present data do not allow to determine the extent to which granted leave is discussed with patients before and after decisions are made. Nevertheless, if patients only passively received the multidisciplinary team’s deliberations and decisions, this could compromise the therapeutic alliance, which is the most important predictor of change in psychotherapy (Flückiger et al., 2018).
Leave applications of patients are first reviewed for their chances of approval. The organizational levels at which this preliminary decision is made vary across hospitals: Applications are optionally checked by the therapist, the ward team, or both. Procedural differences in this regard have been described earlier (Barlow & Dickens, 2018; Neumann et al., 2019). Similarly, different professional groups are involved in the preparation of decision-making body for the granting of leave; in two-fifths of Bavarian forensic-psychiatric hospitals, the nursing staff is organizationally responsible. The importance of nurses in the organizational and content preparation of a decision about granted leave has already been noted (Barlow & Dickens, 2018) and may be related to their expertise in patients’ daily lives (Haines et al., 2018). Due to this expertise greater nurse involvement may also indicate indirect patient participation in terms of a shared decision-making policy during decision-making about granted leave (Selvin et al., 2021). Hospitals also differ in the number of levels at which pre-decisions on granted leave are made within the multi-level process as well as in the professional diversity of the decision-makers involved. For example, for high-risk patients (murderers, arsonists, etc.) some hospitals make decisions in meetings held across wards, thus involving clinicians not directly involved in the treatment. Others include only clinicians who are closely involved in the treatment. The final decision to grant leave is, by law, the responsibility of the chief psychiatrist, which was affirmed by the majority of the participants. In summary, the results show that decisions on granted leave, in line with the legal requirements, are mainly based on a multi-professional vote within a multi-level process. Multidisciplinary work is common practice in forensic settings because it is believed that incorporating different professional perspectives can improve risk management and objectify decisions (Mason et al., 2002). However, the present results do not shed light on the impact of stakeholders’ expertise or any power dynamics on decisions (Haines et al., 2018).
There were two process characteristics required by law that seem to receive little attention. First, the periodic review of a patient’s indication for a granted leave. Given the legal requirement for a review of the treatment plan every six months, this finding was surprising. Because granted leave is considered an essential component of forensic-psychiatric treatment it should reliably be included in treatment planning revisions. Second, only a small proportion of professionals reported the use of a risk assessment approach that distinguishes between the risk of reoffence and absconding during granted leave. In fact, a differential risk assessment seems to take place at only one given hospital. This must be viewed critically. There is strong evidence that only an appropriate risk assessment delivers valid results (Bonta & Andrews, 2016; Hanson & Morton-Bourgon, 2004). However, very few risk assessment tools for granted leave outcomes have been developed and validated thus far (Booth et al., 2021; Hearn et al., 2012; Hilterman et al., 2011; O’Shea & Dickens, 2014; Watson & Choo, 2021). This could mean that even when professionals use valid risk assessment tools (as recommended by Kennedy, 2022), they do so inappropriately because the tools are typically intended to assess a specific risk (e.g., for violence or sexual violence). The missing risk differentiation might also result from a lack of hospitals’ clear definition of misconduct during granted leave or absconding: Only two-fifths of the respondents confirmed a documented definition and handling of absconding or reoffending during granted leave and the associated policy. Bavarian legal requirements hold back regarding definitions and policies for the internal handling of absconding as well.
Notwithstanding the legal requirements, only half of all respondents reported a procedure for revocation granted leave based on documented in-hospital agreements, and only in one hospital was this mentioned by the majority. This result could be a consequence of a lack of legal requirements, as they clarify the circumstances for suspending or revoking a granted leave but do not provide any formal instructions for the procedure. At the same time, however, documented in-hospital agreements for granting leave were reported in 10 of 12 hospitals, even though the legal requirements already govern many procedural aspects for granting leave. Thus, hospitals appear to be attempting to develop and implement their own care models beyond legal requirements. This is an important finding considering that clinical guidelines or related concepts such as recommendations or expert advice positively influence the quality of care and save resources (Gatta et al., 2019).
However, the development of individual in-hospital agreements might inevitably lead to different approaches across hospitals, as the present results show. Cross-hospital standards of excellence for granting leave and revocation of granted leave in forensic psychiatry could improve transparency and coherent decision-making and facilitate consistent crisis and risk management planning (Kennedy, 2022; Vogel et al., 2004). Legal requirements can provide valuable guidance in this regard, but by their nature focus primarily on the safety interests of the public. Because clinicians in Germany are closely involved in the patients’ treatment while having responsibility for decisions about granted leave, they are facing the dilemma between patient rights and community rights (Birgden, 2008; Völlm et al., 2018). Thus, more detailed guidelines on decision-making about granted leave might be needed to help them make decisions more confidently. Such guidelines could be developed by a working group with representatives from all forensic-psychiatric hospitals to increase their acceptance. They should focus, for example, on a uniform structured approach to risk assessment that would give clinicians more confidence in making decisions. This should include the use of valid risk assessment tools that distinguish between (sexual) violence and absconding risk. In addition, such an approach should consider the context of recidivism during granted leave and assist clinicians in risk formulation, as is common in structured professional judgment (Baird & Stocks, 2013; Douglas, 2019). However, more specifically formulated standards of excellence require further empirical evidence on the relationship between decisions about granted leave and recidivism, absconding, or leave rates.
Limitations and future directions
Given the national and international differences in legal requirements for forensic-psychiatric treatment, it is hardly possible to infer the results to decision-making processes for granted leave at hospitals outside Bavaria. Methodologically, the present results only allow limited conclusions about the actual processes in the hospitals because even if the study participants are representatives of all professional groups, their statements only reflect their perception of the process depending on their participation. Therefore, the lesser the involvement of interview participants in a process step, the lesser they can comment on the criteria, and the more biased the overall results would be. In addition to the assessment of the importance of specific characteristics of a decision-making process and the reliability of the resulting decisions, it would be helpful to determine whether there is a relationship between those characteristics and officially recorded data on absconding with or without reoffending.
Conclusion
In summary, the results show that the decision-making processes for granting leave at the Bavarian forensic-psychiatric hospitals follow the legal requirements to a great extent. Nevertheless, this seems to differ across the hospitals and there is a clear improvement potential, especially regarding risk assessment during decisions about granted leave. A common arrangement would be desirable for handling absconding and reoffending during granted leave, risk assessment approaches, and revocation of granted leave. A consistent, transparent, and comprehensible decision-making process could prevent fear-motivated restrictive assessments (Dünkel et al., 2018), minimize restrictions on the freedom of patients (Völlm et al., 2018), and thus address both patient and public safety needs.
Footnotes
Conflict of interest
The authors have no conflicts of interest to report.